Mercor AI jobs and platform overview
AI talent marketplace for credentialed professionals
Mercor matches domain experts — lawyers, physicians, finance professionals, engineers — with paid AI evaluation and training projects for AI labs. Screening is AI-assisted and emphasizes verifiable credentials and written evaluation quality.
Sidequest tracks the Mercor opportunities currently published and lists 268 open Mercor roles right now, with advertised rates running a median of $78/hr across 231 hourly listings. Any rate shown is the figure a listing advertises, not earnings, and availability changes as Mercor opens and closes projects.
Mercor at a glance
Figures reported by Mercor and press coverage, not measured by Sidequest. Company announcements and press reporting, late 2025.
How screening works
CV upload, credential verification, then an AI-led video interview (~20 minutes). Shortlisted candidates receive project-specific onboarding, often with a paid trial task batch.
Who fits best
Credentialed professionals with real domain depth who communicate well in structured written English. Listings frequently ask for licences, bar membership, CFA/CPA or advanced degrees.
Open Mercor roles
All roles →
3D Artists & Video Editors (Must own MacBook)
Complete and narrate real 3D or video production workflows to help an AI lab benchmark and improve its models. The work is part time at approximately 10-20 hours per week and requires your own MacBook or Mac running macOS 15 or higher.
Statistical Physics and Thermodynamics Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and Mercor is recruiting PhD physicists to create, solve, review or audit problems in five narrowly defined statistical-physics areas. Matching is by publication record on the specific phenomenon — not by general familiarity with the methods — and every paper you list is checked against the public record.
Audio Engineer (Speech / TTS Audio Specialist) – Brazilian Portuguese
Edit and quality-check Brazilian Portuguese voice recordings for speech and text-to-speech training datasets. The work combines audio restoration, technical compliance checks, linguistic judgment and consistent batch processing.
Clinical Mental Health Expert — AI Conversation Evaluation
Mercor is recruiting licensed and clinically trained mental health professionals to judge which of two simulated therapy-style conversations reads as more authentic. The task itself is short — roughly 30 minutes, fully remote and asynchronous — with an observed rate of $100/hr and the possibility of being tapped for later clinical evaluation work.
Application Users - macOS Sonoma 14.5 Expert - Professionals
Mercor is recruiting daily macOS Sonoma 14.5 users to produce screen-captured task demonstrations and judgments about how the Mac desktop actually behaves — Finder, System Settings, Spaces, window management and bundled apps. You must already run Sonoma 14.5 as your primary desktop OS and be able to capture screenshots above 2.5 megapixels; no hardware, VMs or licenses are provided.
Stochastic Biophysics Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel covers one narrow area: stochastic autocatalytic growth — chemical master equations, multi-type branching processes and moment hierarchies. You would create, solve, review or audit problems in that area, with written reasoning complete enough for another specialist to verify independently.
Neurologists (MD/DO) – Paid Research Study (Migraine)
Mercor is recruiting US-licensed neurologists who actively treat adult migraine patients to complete a single ~30-minute online survey about their real-world prescribing and management decisions. The observed rate is a one-time $150 payment on verified submission — this is a one-off research study, not ongoing evaluation work.
Application Users - DaVinci Resolve on macOS - Generalist
Mercor is recruiting working colorists, editors and post-production professionals who use DaVinci Resolve on their own Mac to produce screen-capture-based task data and evaluate AI output on real Resolve workflows. You supply the machine, the licence and the display resolution; the platform supplies the tasks and pays an observed $30–40/hr.
Application Users - Word on macOS - Professionals
Mercor is recruiting professionals who use Microsoft Word on macOS in real daily work — long-form drafting, styles, referencing, track changes and review — to produce screen-captured task demonstrations that train and evaluate AI agents on desktop software. You supply your own Mac, your own licensed Word install, and a display capable of screenshots above 2.5 megapixels; observed pay for this band is $60–70/hr.
Application Users - Premiere on Windows - Generalist
Mercor is recruiting working video editors who use Adobe Premiere Pro on their own Windows PC to produce screen-captured task data for AI training and evaluation. You supply the licence, the machine, and a display that can capture screenshots above 2.5 megapixels; observed pay sits in the $30–40/hr range.
Safeguards Analyst / IAEA or Euratom Inspector
Mercor is recruiting nuclear safeguards practitioners — IAEA, Euratom or state-system inspectors, accountancy and instrumentation specialists — to write adversarial prompts and grade how frontier models handle them. The work sits on the dual-use line: a model that refuses a routine material-balance question fails as badly as one that explains how to hide a diversion. Observed pay is $65–75 per task.
AWS Serverless & Infrastructure-as-Code Task Auditor
Audit AWS serverless and infrastructure-as-code tasks used to train and evaluate a frontier lab's models, judging architectural soundness, IaC fidelity, and cross-service integration correctness. Written rubric-based feedback is the deliverable, so you spend as much time explaining why something fails as finding it.
Electrical Engineer
Evaluate and create electrical engineering problems, reference solutions, and scoring rubrics used to improve large language models. The work calls for substantial industry ownership in circuits, power systems, embedded hardware, or related electrical engineering domains.
Materials Science Domain Expert
A full-time W-2 role through Cincinnatus LLC, embedded with a frontier AI lab's research team to improve how models reason about materials science and materials engineering. You write instruction specs and golden solutions, review task and model-output quality, and design benchmarks — on-site in the Bay Area multiple days a week.
Application Users - EViews on Windows - STEM
Mercor is recruiting econometricians and applied economists who use EViews daily on a Windows PC to produce recorded workflow demonstrations and evaluate AI performance on time-series and forecasting tasks. You supply your own licensed EViews installation and a display capable of screenshots above 2.5 megapixels; no software, VMs, or hardware are provided.
Applied Mathematics Benchmark Specialist
Write and verify graduate-level multiple-choice mathematics problems — each with nine subtly wrong distractors, a markdown chain-of-thought solution, and academic citations — for use as AI benchmark items. Work is asynchronous and remote, with an expected commitment of 10+ hours per week at an observed $61–77/hr.
Finance Domain Expert — AI Training & Evaluation
A senior finance practitioner works alongside a frontier AI lab's research team to write instruction specs, produce golden solutions, and QA finance knowledge-work tasks. It is a full-time W-2 role through Cincinnatus LLC, remote day-to-day but requiring Bay Area residency and occasional on-site days.
Pharma Commercial Forecasting Expert — Launch Curves
Mercor is running a paid pilot with a biotech/pharma research team to build and critique rubrics for evaluating commercial drug forecasts and development programs. You'll assess launch curves — uptake, peak share, trajectory, LOE — and judge whether a forecast's underlying assumptions hold up.
Application Users - MATLAB on Windows - STEM
Mercor is recruiting engineers and scientists who use MATLAB daily on a Windows PC to produce screen-captured task work for AI training and evaluation. You supply your own licensed MATLAB install and a display that can capture above 2.5 megapixels; no licences, VMs, or hardware are provided.
Emergency Response Specialist
Mercor is recruiting radiological emergency planners and responders to write single-turn prompts that sit on the dual-use line — benign, dual-use, adversarial — and judge whether frontier models handled them correctly. Every verdict needs a written reference answer and technical rationale a non-specialist can follow, all built strictly from unclassified, non-export-controlled material.
Oncologist
Mercor is seeking actively practicing US medical oncologists with direct clinical trial experience to evaluate and challenge AI reasoning. The work focuses on trial eligibility, treatment decisions, clinical accuracy, and the safety of model-generated oncology recommendations.
Sales Representatives, Wholesale and Manufacturing, Technical and Scientific Products
Mercor is contracting B2B technical sales professionals to grade AI-generated sales artifacts — prospecting emails, discovery calls, proposals, objection handling — for technical accuracy and commercial credibility. The project runs roughly 4–6 weeks at 20+ hours per week, with observed rates in the $70–110/hr range once you clear the first approved task.
English AI Safety Specialist
Review short English-language conversations and label how well AI responses address user requests. Work remotely across assigned batches using provided guidelines and structured feedback formats.
Voice Actor: CX Agent Voice Cloning (USA)
Record clean, expressive American English voice samples that will train and evaluate a research lab's text-to-speech system, including a customer-experience AI agent built on a clone of your voice. Work is remote from your own booth or treated room, roughly 5–10 hours per week, at an observed rate of $50/hr.
Atomic Layer Deposition (ALD) Experts
Mercor is contracting ALD and thin-film process experts to build and grade the scientific data a frontier lab uses to train models on semiconductor processing. Work is remote, US-based, 10–40 hours per week, with pay observed at $84/hr.
Senior Product Designer, UI/UX
A leading AI lab is hiring senior product designers and design engineers to define what "great" UI/UX means for its models and to build interactive web prototypes that demonstrate it. The engagement is W-2 through Cincinnatus LLC at 40 hours per week on weekdays, with observed rates of $80–110/hr.
Application Users - Android Studio on macOS - Code
Mercor is recruiting Android developers who already run Android Studio on their own Mac to produce screenshot-based task and workflow data from real IDE use. You supply the hardware, the licence-free tooling, and a display that captures above 2.5 megapixels; the platform supplies the task specifications and pays an observed $55–65/hr.
Voice Actor: CX Agent Voice Cloning (German)
Record German-language voice samples that will train and evaluate a customer-experience AI agent's text-to-speech voice, working from your own studio setup. Requires native German fluency, Germany-based residency, professional voice work experience, and explicit consent to voice cloning for this single client use case.
Finance Expert, German (Germany) - AI Training
Create German finance tasks, prompts, files and grading rubrics used to train and evaluate AI agents. The work draws on at least five years of experience in corporate finance, banking, investment or audit, plus fluency in German financial practice.
Voice Actor: CX Agent Voice Cloning (Spanish - Latin America)
Record scripted and conversational Spanish audio that will be used to train and clone a voice for a client's internal customer-experience AI agent. Requires native Latin American Spanish with neutral, region-light delivery, a near-professional home recording setup, and explicit comfort with voice cloning.
Mechanical Engineer
Practicing mechanical engineers review AI-generated engineering work, write instruction specs and reference solutions, and build evaluation sets across design, thermal/fluid, stress, manufacturing, and controls. It is a full-time W-2 engagement through Cincinnatus LLC, remote within the US, running 2–3 months initially with possible extension.
Civil Engineers & Architects (Must own MacBook)
Complete and narrate civil engineering or architectural design and drafting workflows to support AI model research. The work requires a compatible Mac, relevant academic or professional experience, and proficiency with Mac-compatible design software.
Computer Power Users & Automation Builders (Must own MacBook)
Complete and narrate multi-app productivity and automation workflows on your own compatible Mac to support AI model research. The part-time project calls for strong macOS fluency, hands-on automation experience, and approximately 10-20 hours of weekly availability.
Atomistic Materials Modeling Expert
Evaluate AI-generated solutions to advanced atomistic materials modelling problems, including DFT, molecular dynamics, energetics and materials property predictions. The work combines technical review, error diagnosis, reference-solution writing and rubric development for scientific AI research.
Administrative Services Managers
Evaluate AI-generated plans, procedures and workflows for facilities, procurement, records management and office operations. The work draws on practical administrative management experience to judge operational soundness, policy compliance and cost-effectiveness.
Data Analysts & Spreadsheet Specialists (Must own MacBook)
Complete and narrate real spreadsheet and data analysis workflows to help evaluate and improve AI models. The role requires an eligible Mac, advanced spreadsheet skills, and approximately 10–20 hours of part-time work each week.
Voice Actor: CX Agent Voice Cloning (Southern USA)
Record scripted and conversational speech in a native Southern American English accent for a text-to-speech system powering a client's internal customer-experience AI agent. Your voice may be cloned for that single use case, and the engagement is short — roughly 5–10 hours per week for one to two weeks at an observed $50/hr.
Certified Medical Coder (ICD-10) — Multilingual Clinical Documentation & AI Evaluation
Review AI-generated clinical notes and real clinical documentation for coding integrity, then write structured feedback engineering teams can act on. Requires an active CCDS, CHC, CCS, or CPC, current work in coding or CDI, US residency, and C1+ fluency in a second language.
Code Plan Reviewer (Plans Examiner) — Vulcan
Author AI evaluation tasks from real permit drawing sets — correction letters, review decisions, marked-up sheets — using the same judgment you apply in a plans examiner seat. Every task you write goes to a second reviewer in your discipline before it ships, and you'll review theirs in turn.
Finance Expert
Mercor keeps a standing pool of finance and accounting professionals who write graded financial problems and score model-produced work for frontier AI labs. It is not a specific opening — you apply once, complete a short AI interview, and get invited to project listings when your specialty comes up.
STEM Expert, Spanish (Spain) - AI Training
Create realistic STEM tasks in Spanish, including workspaces, files, prompts and grading rubrics, to train and evaluate AI agents. The work is remote and suited to scientists or engineers with at least five years of research or industry experience and a degree from a top-20 university in Spain.
Gravitation, Cosmology and Astrophysics Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel needs physicists who have published on torsionful gravity, Nieh-Yan or Chern-Simons inflation, or photohadronic blazar emission to create, solve, review or audit those problems. Assignments are agreed after a narrow subfield match, run about 10 hours a week for 8–10 weeks, and every deliverable must be complete enough for another specialist to verify independently.
Legal Expert, Spanish (Spain) - AI Training
Create realistic Spanish legal tasks, supporting files, prompts and grading rubrics for training and evaluating AI agents. The work draws on at least five years of practice in Spain and deep knowledge of Spanish civil, commercial or employment law.
Software & Firmware Engineering Experts
Mercor is hiring embedded, FPGA/RTL, flight software, and hardware test automation engineers to evaluate AI-generated work on systems that touch real hardware. You'll review firmware, RTL, and test infrastructure for correctness under timing, memory, concurrency, and fault-handling constraints, and write reference problems where the failure modes are physical, not theoretical.
Clinical Survey Research & Methodology Expert (Physician / PhD)
Review and develop physician-facing healthcare surveys for clinical accuracy, methodological quality, and relevance to real-world practice. The work combines therapeutic-area expertise with hands-on healthcare market research and questionnaire design experience.
UI/UX Designers (Must own MacBook)
Complete and narrate real UI/UX workflows on your own compatible Mac to help evaluate and improve AI models. The role is designed for UI/UX designers with 2-3 years of professional experience or current or recent postgraduate HCI or design study.
STEM Researchers - MSc/PhD (Must own MacBook)
Complete and narrate research workflows that help an AI lab benchmark and improve its models. The role is for current or completed STEM MSc or PhD researchers who own a compatible Mac and can contribute approximately 10 to 20 hours per week.
[LONG PROJECT] Anonymization Writer
Review business documents and structured records to identify, type and consistently group personally identifiable information. This contract role requires prior AI data work, careful rule application and a commitment of at least 30 hours per week.
Customer Experience Operations (India)
Own the expert experience across onboarding, support, payments and project matching as a full-time member of Mercor's internal team. Coordinate Engineering, Operations, Product and Talent Success to identify friction, prioritize fixes and keep cross-functional work on schedule.
Finance Expert, Mandarin Chinese (mainland China) - AI Training
Create realistic mainland China finance tasks, workspaces and grading rubrics used to train and evaluate AI agents. The role requires at least five years of relevant experience, Native or fluent in Mandarin Chinese, and detailed knowledge of Chinese financial practice.
Senior Software Engineer, Backend (Node.js, TypeScript, Rust)
A full-time (40 hrs/week) W-2 backend engineering placement through Cincinnatus LLC, embedded in a large enterprise client's consumer platform team. You'd build and operate Node.js, TypeScript, and Rust services behind high-traffic web and mobile experiences, including API layers and AI/ML integrations.
Audiobook QA Expert — Spanish (US)
Listen to AI-narrated Spanish audiobooks in US/North American Spanish and log where the synthesized narration breaks — mispronunciations, skipped or added words, misread numbers, wrong abbreviation expansions, flat intonation, audio artifacts. Roughly 20 hours a week, remote, with each issue categorized and time-spanned in the platform's review tool.
Pediatric Emergency / Urgent Care Physician (MD)
Review and annotate pediatric emergency or urgent care documentation to assess the accuracy, completeness, and safety of AI-generated clinical summaries. This remote, part-time role is for US-based MDs with pediatric acute-care experience and at least 10 hours of weekly availability through mid-November.
Quantum Information Science Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel covers twelve narrowly defined quantum information topics — from GHZ metrology and SU(1,1) interferometry to Weingarten calculus and contraction coefficients. You write, solve, review or audit problems in the subfield you have actually published on, with written reasoning complete enough for another specialist to verify independently.
Apple Watch Users: Earn $60
Mercor is assembling a de-identified research dataset that pairs Apple HealthKit data with linked electronic health records from a fixed list of US health systems. You contribute one Apple Health export that includes a connected provider record with immunizations, labs, medications, vitals, and clinical notes, and consent to share four demographic attributes, for a one-time $60 payment as observed on the platform.
Cardiologist – Clinical Consultant (ATTR Cardiac Amyloidosis)
Review clinical charts to assess the likelihood of transthyretin cardiac amyloidosis and document the evidence behind each assessment. This part-time consulting role supports AI model improvement through consistent labeling, differential diagnosis, and clear communication of uncertainty.
Consulting Expert, German (Germany) - AI Training
Create consulting workspaces, prompts and grading rubrics that reflect strategy, operations and restructuring engagements in Germany. This remote pilot is intended for experienced consultants who are native or fluent in German and have worked with DACH clients, including Mittelstand and industrial companies.
Language & Audio Expert
Mercor keeps a standing pool of linguists, translators, voice actors, and audio professionals who grade and write training data for frontier AI models. Joining is a background check plus a short AI interview, and project invitations arrive later when a client needs your language or specialty.
Human Resources Expert, Spanish (Spain) - AI Training
Create realistic Spanish HR tasks, workspaces and grading rubrics used to train and evaluate AI agents. The work draws on at least five years of HR experience and knowledge of employment law, collective agreements, Social Security and payroll in Spain.
Consumer & Lifestyle Expert
Mercor keeps a standing pool of consumer and lifestyle specialists who write realistic buyer questions, author reference answers, and grade AI model recommendations for accuracy, practicality, and cultural fit. It is not a specific opening — you complete a short AI interview, get verified, and are invited to project listings when a client needs your category.
Fluid Dynamics and Convection Stability Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel covers two narrow areas: linear stability of Rayleigh-Bénard convection with mixed boundary conditions, and Rayleigh-Darcy convection in saturated porous layers. You would create, solve, review or audit problems in your own subfield, with written reasoning complete enough for another specialist to verify independently.
Generalist Expert
Mercor keeps a standing generalist pool for AI training and evaluation projects that need careful reading and general reasoning rather than a specific credential. You write prompts and reference answers, grade model output for correctness and instruction-following, and explain in writing exactly why an answer fails.
Life, Physical & Social Science Expert
Mercor keeps a standing pool of graduate-trained scientists who write research problems, grade model-produced analysis, and flag results that look coherent but would fail peer review. Projects in this pool have posted at $60–120/hr, set per project by scope and depth; joining the pool is not itself a job offer.
Application Users - AutoCAD Mechanical on Windows - STEM
Mercor is recruiting working mechanical drafters and designers who already run AutoCAD Mechanical on their own Windows PC to produce recorded application workflows for AI training and evaluation. Observed pay is $45–55/hr; no licences, VMs, or hardware are supplied, and your display must support screenshots above 2.5 megapixels.
Fire & Life Safety Reviewer — Vulcan
Author fire and life safety evaluation tasks from real drawings, submittals and site photos — egress analyses, correction lists, plan review decisions — with a peer in your discipline checking each one before it ships. Mercor's Vulcan project wants people who currently sit or recently sat in the review seat: fire marshal, fire prevention plan reviewer, fire protection engineer, or sprinkler/alarm designer-reviewer.
Application Users - PyCharm on macOS - Code
Mercor is recruiting working Python developers who use PyCharm daily on macOS to produce screenshot-based task and workflow data for AI training and evaluation. You supply your own Mac, your own PyCharm install, and a display that captures above 2.5 megapixels — no licenses, VMs, or hardware are provided.
Nuclear Engineer, Fuel Cycle
Mercor is recruiting fuel-cycle practitioners to write single-turn prompts that sit on the dual-use line — benign, dual-use, and adversarial — and judge whether frontier models handled each correctly against a written policy standard. Every judgment ships with a reference answer and a technical rationale a non-specialist can follow; observed pay is $65–75 per task.
Architect — Design Coordination (Vulcan)
Author evaluation tasks from real drawing sets, RFIs and site photos, producing the same deliverables you'd issue at work — coordination comments, marked-up sets, RFI responses — plus the defensible correct answer. A second architect in your discipline back-checks every task before it ships.
Investment Banking Expert
Mercor keeps a standing pool of investment bankers who grade and correct AI-generated models, comps, and deal materials for frontier AI labs. You build the artifact yourself — a three-statement model, an LBO, a comp set — then write the rubric that says exactly where the model's attempt fails and why.
iPhone Users: Earn $60 for Sharing Your Fitness Data (US Only)
Mercor pays a one-time $60 for a genuine Apple Health export that combines your on-device HealthKit history with clinical records linked from one of roughly 60 named US health systems. The submission is rejected unless the linked record includes immunizations, labs, medications, vitals, and clinical notes, and all four demographic attributes are consented.
Radiation Safety Officer
Mercor is recruiting named radiation safety officers to red-team frontier models on radiological dual-use questions — writing prompts across benign, dual-use and adversarial tiers, then judging whether the model's response met a defined policy standard. Every judgment needs a written technical rationale a non-specialist can follow; observed pay is $65–75 per task.
Data Analysis Expert
Mercor keeps a standing pool of analysts, data engineers, statisticians, and ML practitioners who write and grade analysis problems used to train and evaluate frontier AI models. You apply once, complete a ~20 minute AI interview, and get invited to specific projects — each naming its own rate, hours, and client — when your specialty comes up.
Application Users - Photoshop on Windows - Generalist
Mercor is recruiting working Photoshop users on Windows to produce screen-captured records of real raster editing, retouching and compositing tasks for AI training and evaluation. You supply your own licensed Photoshop install and a display setup capable of screenshots above 2.5 megapixels — no software, VMs or hardware are provided.
Application Users - Visual Studio Code on macOS - Code
Mercor is recruiting working developers who use Visual Studio Code on macOS daily to produce screen-captured task demonstrations and judgments that train AI models on real editor workflows. You must already own the Mac and a display that can capture screenshots above 2.5 megapixels — no hardware, licences, or VMs are provided.
Customer Success - Operations (India Region)
Full-time support and operations work on Mercor India's Talent Success team, answering questions from experts applying to roles, taking AI-led interviews, and getting paid. You triage bug reports with engineering, write talent-facing documentation, and push recurring issues back into the product instead of resolving them manually forever.
Inpatient Hospitalist (MD) — Clinical Documentation & AI Evaluation
Mercor is recruiting practicing US inpatient hospitalists (MD) to review hospital documentation and evaluate AI-generated H&Ps, progress notes, and discharge summaries. Work is remote and part-time, scheduled around an active clinical practice, with an observed rate of $170/hr.
Synthetic / Process Chemist
Mercor is recruiting practising synthetic and process chemists to write adversarial prompts and grade frontier-model responses on misuse judgment. The work is paid per task in an observed $65–75 band, is fully remote and asynchronous, and is heavily writing-based — every verdict needs a rationale a non-chemist can follow.
SWE-Bench Task Auditor
Audit repository-level software-engineering benchmark tasks — reference patches, test harnesses, and Docker-based grading — for correctness, reproducibility, and leakage. Deliver rubric-based written verdicts that a frontier lab can act on when deciding whether a task enters a training or evaluation set.
Customer Success Engineer (India)
Front-line technical support for Mercor's talent platform, investigating bug reports from thousands of weekly users and debugging across an AI + SaaS stack using logs, network inspection, and database queries. Remote from India, six days a week, with substantial overlap with Pacific Time.
Arts & Design Expert
Mercor keeps a standing pool of working designers, writers, and creative professionals who grade AI-generated design, prose, and imagery on craft — and write the briefs and reference answers models are graded against. It is not a specific opening: you complete a short AI interview, get verified into the arts and design pool, and are invited to individual projects when one matches your discipline.
Preclinical R&D Expert – In-house Drug Development
Mercor is recruiting preclinical R&D scientists from in-house drug developers to review, annotate, and stress-test model outputs for an AI lab building physics-based foundation models for chemistry and drug discovery. Work is remote and hourly, with direct exposure to the client team and observed rates of $60–100/hr.
Voice Actor: CX Agent Voice Cloning (African American)
Record scripted and conversational voice samples that will be used to train and clone a single customer-experience AI agent voice. Requires a USA-based voice actor with natural African American English delivery, a near-professional home recording setup, and explicit consent to voice cloning for this internal use case.
Employer Group Benefits Brokers & Consultants – One-Time Paid Research Survey
Complete a one-time online research survey about advising US employers on group medical, pharmacy or retiree benefits. The survey takes approximately 20 to 23 minutes and combines multiple-choice questions with short voice-recorded responses.
Finance Expert, Spanish (Spain) - AI Training
Create realistic Spanish finance tasks, workspaces and grading rubrics used to train and evaluate AI agents. This remote role requires at least five years of relevant experience, native or fluent Spanish, and knowledge of financial practice in Spain.
Public Defenders — Paid Research Study
Mercor is recruiting 100 current or former U.S. public defenders to document, in a structured survey, the actual tasks their caseload work consisted of and how those tasks depend on one another. Wave 1 pays $100 per reviewed hour up to 90 minutes ($150 cap), with a possible later survey reviewing a combined task tree at terms confirmed separately.
Voice Actor: CX Agent Voice Cloning (Swiss German)
Record Swiss German voice samples that will be used to train and clone a voice for a client's internal customer-experience AI agent. Requires Switzerland-based native speakers with voice acting or broadcast experience and a near-professional home recording setup.
Condensed Matter Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel recruits condensed matter PhDs to write, solve, review or audit problems inside nineteen narrowly defined subfields. Matching is by publication on the specific phenomenon — not condensed matter broadly — and every assignment requires written reasoning another specialist could verify independently.
Behavioral Health Expert, AI Safety and Model Evaluation
Evaluate AI responses to sensitive conversations about relationships, emotional wellbeing, personal beliefs and difficult decisions. You will identify unsafe or unbalanced patterns, write evidence-based rationales and help develop evaluation rubrics and test scenarios.
Brazil Personal & Cross-Border Finance – Paid Research Study
Mercor is recruiting Brazil-based participants with real financial ties to the United States for a ~20-minute online survey on banking, spending, and cross-border money management. Compensation is observed at $100 one-time, paid after the submission is verified.
Cybersecurity Research Expert – Offensive Security & Vulnerability Research
Review AI-generated vulnerability analyses, exploit chains, and security writeups for technical correctness and real exploitability, then document where the reasoning breaks down. Mercor screens for verifiable offensive security credentials — CVEs, top-tier CTF placements, accepted bug bounty findings, or published security research.
Get $60 for Supporting Wellness Research (US Only)
Mercor is assembling a de-identified health dataset by pairing Apple HealthKit data with linked clinical records from a fixed list of US health systems. You contribute one export from your own iPhone Health app — including immunizations, labs, medications, vitals, and clinical notes — for a stated $60 one-time payment.
Mechanical Engineering Experts
Mercor is recruiting practicing mechanical engineers — structures/stress, thermal, mechanical design, or dynamics and controls — to review AI-generated engineering work and write reference problems grounded in real hardware. The work is remote and flexible, with observed rates of $100–120/hr for engineers who have personally owned hardware from analysis through qualification.
Private Equity Associate – Real Estate (US)
Build a full real estate transaction model from a synthetic deal document pack with no template provided, then peer-review two anonymized models from other contractors against a fixed scorecard. The engagement runs roughly 20–25 hours over 1.5–2 weeks, and generative AI may not be used for either the build or the reviews.
Graphic Designers (Must own MacBook)
Complete and narrate real graphic design workflows on your own Mac to help an AI lab benchmark and improve its models. This part-time Mercor project expects approximately 10-20 hours per week and requires a Mac running macOS 15 or higher.
Medical Expert, German (Germany) - AI Training
Experienced physicians will create realistic German medical tasks, prompts, workspaces, files and grading rubrics for AI training. The role requires at least five years of clinical practice, German Approbation and native or fluent German.
Private Detectives and Investigators
Review AI-generated investigation plans, research summaries, surveillance reports, and case analyses for accuracy, rigor, and legal and ethical compliance. This short-term project requires at least four years of relevant investigative experience and availability for 20 or more hours per week.
Mathematical Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel recruits mathematical physicists to write, solve, review or audit problems inside one of four narrow areas — special functions, integrable systems, permutation combinatorics for random tensor networks, or Fefferman-Graham conformal geometry. Matching is done on your publication record for the specific phenomenon, not the adjacent one, and the standard of written proof sits closer to mathematics than to physics.
Urologists & Medical Oncologists (Prostate Cancer) – Paid Research Study
Mercor is recruiting US board-certified urologists and medical oncologists who actively manage metastatic prostate cancer to complete a single ~30-minute online survey on systemic treatment decision-making. Compensation is observed at $175 one-time, paid after the submission is verified.
Engineering Expert
Mercor keeps a standing pool of practicing engineers who write design problems, grade model-generated analysis, and flag where a technically correct calculation would fail in the shop, the field, or the code review. It is not a specific opening — you join the pool, and invitations to funded projects arrive when a client needs your discipline.
Electrical Engineering Expert
Mercor keeps a standing pool of electrical engineers who write circuit and systems problems for AI models to attempt, then grade the model's designs against a reference answer. Work is project-based and intermittent — past rates have run $70–120/hr, set per project by scope and depth.
Private Equity Associate – Consumer & Leisure (France/Europe)
Build a full LBO-style financial model of a synthetic consumer/leisure transaction from a raw document pack with no template, then peer-review two anonymized models from other contractors against a fixed scorecard. Roughly 20–25 hours over 1.5–2 weeks, with no generative AI permitted at any stage of building or reviewing.
Property, Real Estate, and Community Association Managers
Evaluate AI-generated budgets, lease analyses, maintenance plans, governance documents, and property management workflows using your real-world operating experience. This short-term project requires at least four years in property or community association management and a commitment of 20 or more hours per week.
Physics Expert
Mercor keeps a standing physics pool that frontier AI labs draw from when they need derivations written, graded, and stress-tested by someone who has actually done the work. You write problems from your own subfield with reference solutions, and grade model attempts on whether the approximations, limits, units, and scaling hold up — not just whether the final expression matches.
Financial Managers
Mercor is contracting finance managers to grade AI-generated budgets, forecasts, financial plans and reporting artifacts, and to rank competing model responses on financial soundness and regulatory compliance. The engagement runs roughly 4–6 weeks at 20+ hours per week, remote and asynchronous, with observed rates of $70–110/hr after a task-based qualifying period.
Materials Science Expert
Mercor keeps a standing pool of materials scientists who write and grade problems used to train and evaluate frontier AI models. Work is project-based and intermittent: you author problems from systems you actually processed and characterized, then judge whether a model's structure–property reasoning and material selections would survive service conditions.
Management Consulting Expert
Mercor keeps a standing pool of strategy, operations, and transformation consultants who write case prompts and grade AI-generated business analyses for frontier model research teams. You are not applying to a dated opening — you pass a ~20-minute AI interview, get verified, and are invited to specific projects when one matches your background.
Voice Actor: CX Agent Voice Cloning (South African English)
Record roughly four hours of scripted South African English speech that will be used to clone a voice for a client's customer-experience AI agent. You must be based in South Africa, speak with an authentic South African accent, have a quiet room and a decent microphone, and be comfortable with voice cloning for this single internal use case.
Electrical Test Automation Engineer
Mercor is recruiting practicing electrical test automation engineers for a short-term project that turns real bench experience into training and evaluation material for AI models. The work is writing and reviewing automated test problems — instrument control, DAQ integration, HIL setups, failure root-cause reasoning — and judging whether model-generated answers would survive contact with an actual test rack.
Engineering & Software Domain Expert
Senior software engineers write the instruction specs, golden solutions, and benchmarks that define correct engineering work for a frontier AI lab, and review task and model output quality against that bar. It is a full-time W-2 role through Cincinnatus LLC, hybrid on-site in the Bay Area, with an initial six-month engagement.
Nuclear Security Engineer
Mercor is recruiting nuclear materials and safeguards practitioners to write adversarial prompts and grade frontier model responses on where the dual-use line falls. Work is per-task, remote and asynchronous, with observed rates of $65–75 per task, and must be built entirely from unclassified, non-export-controlled knowledge.
Application Users - Vivado on Windows - STEM
Mercor is recruiting FPGA and RTL engineers who already run AMD/Xilinx Vivado on their own Windows PC to produce screen-captured workflow data for AI training and evaluation. No licenses, VMs, or hardware are provided, and your display must support screenshots above 2.5 megapixels.
Application Users - FruitLoops Studio on Windows - Generalist
Mercor is recruiting working FL Studio (FruitLoops Studio) users on Windows to demonstrate real production workflows and judge how AI models handle DAW tasks. You work in your own installed copy, capture high-resolution screenshots of your screen, and document what you did and where a model's answer would break down.
Application Users - PowerPoint on Windows - Professionals
Mercor is recruiting working professionals who use PowerPoint on Windows daily to produce screen-captured task demonstrations that train and evaluate AI models on real desktop workflows. You supply your own licensed PowerPoint installation and a display capable of screenshots above 2.5 megapixels; no software, VM, or hardware is provided.
Application Users - Stata SE on Windows - STEM
Mercor is recruiting working Stata SE users on Windows to produce screen-captured task demonstrations and evaluation data for AI training. You must already own a licensed Stata SE install on a Windows PC and a display setup capable of screenshots above 2.5 megapixels — no software, VMs, or hardware are provided.
Legal Domain Expert — AI Training & Evaluation
A full-time W-2 role through Cincinnatus LLC, embedded with a frontier AI lab's research team to build legal evaluation sets, instruction specs, and golden solutions. Requires a JD, active U.S. bar admission, 5+ years of senior-level practice, and Bay Area residency with occasional on-site days.
Source Security Specialist
Mercor is recruiting radiological source security practitioners to write adversarial prompts and grade frontier model responses on Category 1 and 2 source misuse questions. Observed pay is $65–75 per task, remote and asynchronous, with every judgment requiring a written rationale a non-specialist can follow.
Medicare Advantage Brokers Survey – Agent & Enrollment Insights
Mercor is paying licensed Medicare Advantage agents and brokers $140 to complete a ~20-minute online survey combining multiple-choice questions with short voice-recorded answers. The study feeds an AI research lab's work on tools for Medicare plan selection, agent workflows, and member experience.
Health Physicist
Mercor is recruiting operational health physicists to write adversarial prompts and grade frontier model responses on radiological safety questions that sit on the dual-use line. Work is remote, task-based, and writing-heavy: every judgment needs a written rationale plus a reference answer with technical reasoning, at an observed $65–75 per task.
Nuclear Engineering & Safeguards Experts for Red Team
Mercor is recruiting nuclear fuel-cycle and safeguards specialists to write adversarial prompts and grade frontier-model responses on misuse potential. Every judgment requires a written rationale a non-specialist can follow, and all work must stay clear of classified or export-controlled material.
Audiobook QA Expert — Portuguese (Brazil)
Listen to AI-narrated audiobooks in Brazilian Portuguese and log every place the synthesized voice slips — misread numbers, wrong stress, skipped words, audio artifacts — marking the exact span where each issue occurs. Roughly 20 hours per week, fully remote, with observed pay in the $15–20/hr range.
PDF Annotation & Transcription Experts – Gujarati
Annotate the structure of real Gujarati PDF pages — bounding and typing every region, setting reading order, and transcribing all text character-for-character in Gujarati script. Every task is reviewed end to end by a second Gujarati expert, and experienced annotators take on that review work themselves.
PDF Annotation & Transcription Experts – Odia
Annotate the structure of real Odia-language PDF pages — bounding and typing every region, setting reading order, linking captions to figures — and transcribe all text character-for-character in Odia script, including handwriting. Every task is reviewed end to end by a second Odia expert, and experienced annotators take on that review work.
Spanish Tutors
Mercor is recruiting experienced Spanish tutors to run recorded 20–30 minute one-on-one lessons with native English-speaking learners for a language-learning platform. The engagement runs roughly three weeks at about seven hours per week, remote, with pay observed in the $60–65/hr range.
Electrical Design Specialist
Mercor is recruiting practicing electrical design engineers to work through complex power-system and electrical design problems for an AI lab's model evaluation project. You'll produce and critique load flow, short-circuit, protection coordination and equipment sizing work, and judge whether model-generated engineering reasoning would survive a real design review.
Kubernetes Task Auditor
Audit Kubernetes training and evaluation tasks for a frontier AI lab, checking that scenarios, manifests, and expected fixes actually hold up against a real cluster. Work is rubric-based: you score each task and write the reasoning that justifies the score.
PDF Annotation & Transcription Experts – Bengali
Annotate the structure of real Bengali PDF pages — bounding and typing every region, setting reading order, and transcribing all text character-for-character in Bengali script. Every task is reviewed end to end by a second Bengali expert, and experienced annotators take on that review work.
Epidemiologist — Patient Population Sizing
Mercor is hiring a US-based epidemiologist to build and critique rubrics that judge how AI systems size patient populations for commercial drugs and development-stage programs. The pilot runs roughly 10–20 hours over one to two weeks, with observed rates of $150–175/hr and the possibility of continuing as the work scales.
Insurance Domain Expert
A full-time, hybrid Bay Area position writing instruction specs, golden solutions, and benchmarks that teach frontier models how insurance and actuarial work is actually done. Hired as a W-2 employee of Cincinnatus LLC and placed inside a leading AI lab's research team, using client-issued accounts and tools.
Pharmaceutical R&D Domain Expert
A senior drug development scientist works on-site with a frontier AI lab's research team, writing instruction specs and golden solutions for pharmaceutical R&D tasks and vetting model outputs for unsupported efficacy, safety, or regulatory claims. This is full-time W-2 employment through Cincinnatus LLC, hybrid in the Bay Area, for an initial six-month engagement.
Application Users - Quartus II on Windows - Code
Mercor is recruiting FPGA and digital design engineers who already run Quartus II on their own Windows machine to produce screen-captured task work that trains and evaluates AI models on real EDA workflows. You supply the licence, the installation, and a display capable of screenshots above 2.5 megapixels; the platform supplies the tasks and pays an observed $55–65/hr.
Application Users - Windows 11 Professional Expert - Professionals
Mercor is recruiting Windows 11 Professional power users to generate and evaluate real desktop workflows — Settings, File Explorer, window management and built-in system tools — as training and evaluation data for AI agents. You must already run Windows 11 Pro as your daily driver and be able to capture screenshots above 2.5 megapixels; no licenses, VMs or hardware are provided.
Securities, Commodities, and Financial Services Sales Agents
Mercor is contracting brokers, investment advisors, and financial services sales professionals to grade AI-generated trade recommendations, suitability assessments, and market commentary. Work is task-based at first, shifting to hourly after your first approved task, at an observed $70–110/hr on a 4–6 week project.
Staff Software Engineer, Mobile (Android, Kotlin, Full Stack)
A full-time W-2 placement through Cincinnatus LLC embedding you in a large enterprise client's mobile product team, shipping Android features in Kotlin and stepping into backend, API, and web work when the feature demands it. Mercor screens for real native Android depth plus at least some production experience outside mobile, and for fluency with AI coding tools as part of normal day-to-day work.
Voice Actor: CX Agent Voice Cloning (Spanish - Peru)
Record clean, expressive Peruvian Spanish voice samples that will be used to train and evaluate a client's internal customer-experience AI agent, including voice cloning. Open to Peru-based female native speakers with a professional or near-professional home recording setup and 5–10 hours of weekly availability.
Primary Care Physician (MD) — Outpatient Clinical Documentation & AI Evaluation
Practicing US primary care physicians review outpatient notes and AI-generated documentation, flagging omissions, hallucinations, and clinical inaccuracies against detailed annotation guidelines. Remote and part-time at a minimum of 10 hours per week for roughly three months, with observed pay in the $170–190/hr range.
Civil / Survey Engineer — Vulcan
Author AI evaluation tasks from real civil and survey documents — grading plan reviews, ALTA Table A items, earthwork quantities, title exception schedules — and supply the defensible correct answer. Every task you write is checked by a second PE or PLS in your discipline before it ships.
Nonproliferation Analyst
Mercor is recruiting nuclear nonproliferation practitioners to write and grade single-turn prompts that test whether frontier models can distinguish a routine export-classification question from a procurement-circumvention one. You write the prompt, judge the model's response against a policy standard, and write the reference answer with the technical reasoning behind it — all from unclassified, non-NDA-encumbered knowledge.
Audio Engineer (Speech / TTS Audio Specialist) – French
Clean, quality-check, and evaluate French-language voice recordings destined for TTS and speech AI training pipelines. Work is batch-based and spec-driven, combining audio restoration in tools like iZotope RX with critical listening judgments about French pronunciation and naturalness.
Medical Expert, Spanish (Spain) - AI Training
Experienced physicians will create and evaluate realistic AI training tasks based on medical practice in Spain. The work includes prompts, supporting files and grading rubrics grounded in SNS protocols and Spanish clinical guidelines.
AI Developer Trace Task Auditor
Audit end-to-end AI-assisted coding sessions — the full trace of prompts, tool calls, edits, and test runs — and judge whether the work was correct, the workflow sound, and the reasoning defensible. You write rubric-anchored feedback that a frontier lab uses to train and evaluate coding models.
Pediatric Inpatient Nurses (RN)
Mercor is recruiting pediatric inpatient RNs to evaluate AI-generated clinical outputs against real nursing flowsheet documentation and to annotate pediatric assessment data for training sets. Work is remote and project-based, with observed rates of $55–65/hr for reviewers with recent bedside acute care experience.
Project Management Specialists
Mercor is contracting project managers to evaluate AI-generated project artifacts — plans, status reports, stakeholder emails, budget and staffing decisions — and rank competing model outputs on planning rigor and execution feasibility. The engagement runs roughly 4–6 weeks at 20+ hours per week, with observed rates in the $70–110/hr range after a first approved task.
Nurse Practitioners
Mercor is contracting experienced nurse practitioners to review AI-generated clinical assessments, treatment plans, and patient education content for accuracy and guideline adherence. The engagement runs roughly 4–6 weeks at 20+ hours per week, remote and largely asynchronous, with observed rates of $120–150/hr after an initial task-based qualification step.
Excel Expert — General
Build and grade realistic Excel tasks — models, dashboards, Power Query pipelines, VBA — used to train and evaluate frontier LLMs on spreadsheet work. W-2 employment through Cincinnatus LLC, US-based only, 40 hrs/week weekdays through end of September and 20 hrs/week after.
Business Operations Expert
Mercor keeps a standing pool of business operations practitioners who write and grade operational reasoning tasks used to train and evaluate frontier AI models. You are not applying to a dated opening — you complete a ~20-minute AI interview, get verified, and are invited to specific project listings when one matches your function.
Chemistry Expert
Mercor keeps a standing chemistry pool that frontier AI labs draw from when they need synthesis and characterization judgment graded by someone who has actually run the reactions. You write problems from your own bench or computational work, grade model-proposed routes and mechanisms, and flag where a paper-plausible answer would fail in the flask.
Application Users - Blender on Windows - Generalist
Mercor is recruiting working Blender artists on Windows to produce screen-captured records of real 3D tasks — modelling, sculpting, animation, rendering — for AI training and evaluation. You supply your own licensed-free Blender install, Windows PC, and a display that captures above 2.5 megapixels; observed pay runs $30–40/hr.
Mechanical Engineering Expert
Mercor maintains a standing pool of mechanical engineers who write design problems, grade model-generated engineering analysis, and flag where a model's answer is arithmetically correct but physically unbuildable. It is not a specific opening — you join the pool, and project invitations with named rates and clients follow when one starts.
Industrial & Manufacturing Engineers (Must own MacBook)
Complete and narrate practical industrial and manufacturing engineering workflows to help benchmark and improve AI models. The work is part time and requires an eligible MacBook or Mac running macOS 15 or higher.
AI Assistant User – Weekly Feedback Study
Use Meta Muse or Instinct across at least three parts of your life, then document and rate the assistant’s proactive messages each week. Participants also submit conversation exports with personal details removed and are expected to continue with the same study group over time.
Consulting Expert, Spanish (Spain) - AI Training
Create realistic Spanish-language consulting tasks, workspaces and grading rubrics for training and evaluating AI agents. The work draws on management or strategy consulting experience with strategy, operations and transformation engagements in Spain.
STEM Expert, Mandarin Chinese (mainland China) - AI Training
Create realistic STEM workspaces, prompts and grading rubrics that reflect scientific and engineering practice in mainland China. Evaluate AI agents using Mandarin Chinese materials, quantitative analysis and technical documentation.
Human Resources Expert, German (Germany) - AI Training
Create German HR workspaces, prompts and grading rubrics used to train and evaluate AI agents. The work draws on substantial HR experience and knowledge of German labor law, works councils, collective agreements, payroll and social insurance.
Molecular Biology Experts
Design nucleic acid construct tasks — primers, plasmids, gRNAs, mRNA constructs, HDR templates — and write the expert solutions that become ground truth for a frontier AI lab's models. W-2 employment through Cincinnatus LLC, minimum 20 weekday hours per week, observed rates of $70–105/hr.
Lean Engineer, Formal Mathematics (Lean 4, Mathlib, Theorem Proving)
Write and review Lean 4 proofs against current mathlib, and formalize natural-language mathematics so the formal statement actually says what the original said. You will also audit AI-generated Lean, identifying proofs that compile but prove the wrong thing, and write specific feedback for lab researchers.
Electrical & Hardware Engineering Experts
Mercor is recruiting hands-on electrical and hardware engineers — RF, power electronics, analog/mixed-signal, PCB design, or signal integrity — to evaluate how AI models reason about real hardware problems. Work is remote and project-based, with observed rates of $100–120/hr for engineers who have personally owned bring-up, debug, and validation on shipped products.
[LONG PROJECT] Anonymization Quality Reviewer
Review anonymization work across business documents and structured records, catching missed personal information, classification errors, and incorrect entity groupings. Apply frequently updated guidelines without either approving defective work or returning work that is already correct.
PDF Annotation & Transcription Experts – Korean
Annotate the structure of real Korean PDF pages — bounding and typing every region, setting reading order, and linking captions to their figures — then transcribe every text region character-for-character in Hangul, including hanja and handwriting. Every task is reviewed end to end by a second Korean expert, and experienced annotators take on that review role.
Mathematics Expert
Mercor keeps a standing mathematics pool that frontier labs draw from when they need proofs graded and problems written at research or competition level. You write problems with reference solutions, mark model-written proofs step by step, and flag answers that are right for reasons that don't actually prove anything.
High Energy Physicist (PhD)
Mercor is recruiting high energy theory and nuclear theory PhDs to create, solve, review or audit research-level physics problems for CritPt, a public benchmark testing whether frontier models can do genuine physics research reasoning. Matching is narrow — you must be able to point to your own published papers on one of seven specific phenomena, from AdS/BCFT end-of-the-world branes to one-loop quasi-PDF matching to dark photon signals in interferometers.
Cybersecurity Practitioner: Paid Expert Interviews (SOC, Incident Response, Detection, AppSec)
Mercor is paying practicing enterprise security professionals for two to three one-hour video interviews about how SOC, IR, detection engineering, and AppSec work actually gets done. The conversations feed the design of a benchmark measuring AI agents on real cyber defense tasks — nothing to build, label, or submit.
Education & Training Expert
Mercor maintains a standing pool of educators who write teaching problems and grade AI-generated explanations, lessons, and assessments for frontier model labs. Joining is a resume submission plus a roughly 20-minute AI interview; paid work comes later, when a project matches your subject and level.
Atomic, Molecular and Optical Physicist (PhD)
CritPt is a public benchmark of research-level physics problems, and this panel recruits AMO physicists to write, solve, review or audit problems inside one of seven narrowly defined subareas. Matching is by publication record on the specific phenomenon — command of the methods alone is not enough.
Office & Productivity Software Expert
Mercor maintains a standing pool of people who work at a professional level in Excel, Word, PowerPoint, Google Workspace, or office scripting, and draws from it when AI labs commission evaluation projects. The work is writing realistic office tasks with reference answers and grading model-generated documents, spreadsheets, and decks for correctness and real-world durability.
Humanities Expert
Mercor keeps a standing humanities pool that AI labs draw from when they need scholarly judgment on interpretation, sources, and argument. You write field-specific questions with reference answers, grade model-written interpretive prose, and document exactly where fluent output misreads a text or fabricates a citation.
Software Engineering Expert
Mercor keeps a standing pool of professional software engineers who write, grade, and stress-test the problems frontier models are evaluated against. Work is project-based and intermittent — you join the pool after a short AI interview and get invited when a client project needs your stack.
Asset & Wealth Management Expert
Mercor keeps a standing pool of portfolio managers, allocators, and private client advisors who grade AI-generated investment work — portfolio construction, allocation logic, tax and withdrawal sequencing, attribution. You write the scenarios and reference answers models are measured against, and flag where a recommendation reads fine on paper but would breach a real mandate. Comparable projects have posted at $100–130/hr, set per project.
Legal Expert (J.D./LL.B.)
Mercor keeps a standing pool of licensed attorneys who draft reference answers and grade AI-generated legal work — motions, contract provisions, privilege calls, citation checks. Joining the pool is not a job offer: you complete a ~20-minute AI interview, get your license verified, and are invited to specific project listings when one matches your practice area.
Audiobook QA Expert — Portuguese (Portugal)
Listen to AI-narrated audiobooks in European Portuguese and log every place the synthesis slips — mispronunciations, skipped words, misread numbers, flat intonation, audio artifacts. Roughly 20 hours per week, remote, with each issue categorized and timestamped to the exact span in the review tool.
Company Analysis Expert — Event Propagation
Map the economic network around a specific company event — suppliers, customers, competitors, substitutes, geographic exposure — and estimate how the catalyst transmits to each related name, working only from evidence available at a stated cutoff. You will also grade model-generated versions of the same analysis against your own standard, at an observed band of $150–250/hr.
Nuclear Medicine Physicist / Cyclotron RSO
Mercor is recruiting nuclear medicine physicists, radiopharmacy specialists and cyclotron RSOs to red-team frontier models on radiological dual-use questions. You write single-turn prompts across benign, dual-use and adversarial tiers, score the model's replies against a policy standard, and write the reference answer with its technical reasoning.
Equity Research Expert — Single-Name Analyst Publications
Mercor is recruiting former sell-side equity analysts to forecast when a named analyst will publish after a catalyst and what that note will say, reasoning strictly from a fixed evidence cutoff. Work includes writing point-in-time forecasts with documented reasoning and grading model-generated equity analyses against your own published standard. Observed pay band: $150–250/hr.
Forensic / Toxicological Chemist
Mercor is recruiting forensic chemists and toxicologists to write adversarial prompts and grade frontier-model responses on chemical misuse questions. The work is judgment on the dual-use line — deciding when a substance-identification or precursor question is routine casework and when it is a synthesis request in disguise — with a written rationale for every call. Observed pay is $65–75 per task.
Biotech Investment Analyst — Drug Asset & Company Assessment
Mercor is running a short paid pilot in which biotech investment analysts build and critique rubrics for evaluating commercial drugs and clinical-stage development programs. You will assess roughly 5–10 companies end-to-end — mechanism, clinical data, competitive position, value drivers — and judge whether an investment-style thesis holds up under scrutiny.
Industrial Hygienist / Process Safety Engineer
Mercor is recruiting industrial hygienists and process safety engineers to write and grade prompts that test whether frontier models can separate legitimate chemical-safety questions from misuse attempts. You'll author benign, dual-use and adversarial prompts, judge responses against a written policy standard, and draft the reference answer with technical reasoning. Observed pay is $65–75 per task.
Analytical Chemist
Mercor is recruiting analytical chemists who have developed and validated trace-level methods for hazardous compounds to red-team frontier AI models on dual-use chemistry questions. You write single-turn prompts across benign, dual-use and adversarial tiers, judge model responses against a policy standard, and write the reference answer with its technical reasoning.
Commercial Blasting Engineer
Mercor is recruiting licensed blasters and blasting engineers to write adversarial prompts and grade frontier-model responses on explosives-related questions. The work is about drawing the dual-use line precisely — a model that refuses a routine vibration-compliance question fails as badly as one that answers a diversion request.
Federal Certified Materials Specialist / Incident Analyst
Mercor is recruiting certified explosives and energetic-materials practitioners to write adversarial prompts and grade frontier model responses on where the dual-use line falls. Work is task-based, remote and asynchronous, with an observed band of $65–75 per task.
Hazardous Device Specialist / Public Safety Professional
Mercor is recruiting certified bomb technicians, military EOD, and post-blast investigators to red-team frontier AI models on the dual-use line in energetic materials. You write single-turn prompts at three risk levels, judge the model's responses against a written policy standard, and draft the reference answer with technical reasoning — observed pay $65–75 per task.
Software Engineer — Agentic Search Systems
Mercor is screening engineers who have owned retrieval or relevance on production search systems, including agentic search built on LLMs. The entry point is a 25-minute conversational interview about how you evaluate search quality — no coding and no take-home.
Specialty Materials Chemist — Pyrotechnics or Propulsion
Mercor is recruiting formulation and synthesis chemists from pyrotechnics, propellant and energetic-materials work to red-team frontier models on dual-use judgment. You write single-turn prompts at three risk levels, score the model's responses against a written policy standard, and draft the reference answer with technical reasoning behind it.
Radio Frequency and Electromagnetic Engineering Expert
Mercor is recruiting RF and electromagnetic engineers to solve and review hard technical problems — antennas, propagation, microwave circuits, EMC — on a short-term project for a technology client. Observed pay is $80–95/hr, with work delivered remotely and asynchronously against project-specific evaluation criteria.
Expert Project Manager
Run day-to-day operations for data projects that train LLMs on web browsing tasks — tracking annotator throughput and quality in spreadsheets, reviewing work samples, and keeping contributors unblocked. Mercor lists this at $50–60/hr as observed on the platform; hours and duration vary by project.
Forensic Scientist — Nuclear Materials
Mercor is recruiting nuclear forensics and safeguards practitioners to red-team frontier models on the dual-use line — writing single-turn prompts labelled benign, dual-use, or adversarial, then judging whether the model's response met a defined policy standard. Every judgment ships with a written reference answer and technical rationale a non-specialist can follow; observed pay is $65–75 per task.
Audiobook QA Expert — Italian
Listen to AI-narrated Italian (Italy) audiobooks and log every place the synthesized voice slips — mispronounced or skipped words, misread numbers, wrongly expanded abbreviations, flat or wrong-sounding prosody, audio artifacts — marking the exact span in the review tool. Roughly 20 hours a week, fully remote, with observed pay in the $15–20/hr range.
Audiobook QA Expert — German
Listen to AI-narrated German audiobooks and log every place the synthesis slips — mispronunciations, skipped words, misread numbers, flat intonation, audio artifacts — marking the exact span in review tooling. Roughly 20 hours a week, remote, with observed pay of $15–20/hr.
Application Users - Excel on macOS - Professionals
Mercor is recruiting finance and operations professionals who use Excel on macOS daily to record real spreadsheet work — modelling, formulas, pivot tables, charting — as training and evaluation data for AI systems. You supply your own Mac, your own Excel licence, and a display capable of screenshots above 2.5 megapixels; observed pay runs $60–70/hr.
Application Users - SolidWorks Premium on Windows - STEM
Mercor is recruiting mechanical and product design professionals who use SolidWorks Premium daily on their own Windows PC to produce screen-captured workflow data for AI training. You supply the licence, the machine, and a display capable of screenshots above 2.5 megapixels; the platform supplies the tasks and pays an observed $45–55/hr.
Application Users - Adobe Illustrator on Windows - Generalist
Mercor is recruiting working Adobe Illustrator users on Windows to produce and document real vector workflows that train and evaluate AI models. You use your own licensed copy of Illustrator and a display capable of screenshots above 2.5 megapixels — no software, VM, or hardware is supplied.
Voice Actor: CX Agent Voice Cloning (UK English)
Record scripted and conversational voice samples in neutral standard UK English for a research lab's customer-experience TTS agent, with your voice explicitly cloned for that single internal use. Requires professional voice work experience, a UK base, and a treated home recording setup capable of clean, consistent takes at 5–10 hours per week.
Application Users - VMware Fusion on macOS - Code
Mercor is recruiting practitioners who use VMware Fusion daily on their own Mac to produce screen-captured task data for AI training. You supply the hardware, the macOS install, and the working knowledge of virtualization; the platform supplies the task structure and pays an observed $55–65/hr.
Application Users - Inventor Professional on Windows - STEM
Mercor is recruiting practising mechanical and product designers who use Autodesk Inventor Professional daily on their own Windows PC to produce screen-captured task work for AI training and evaluation. You supply the licence, the machine, and a display capable of screenshots above 2.5 megapixels; the platform supplies the tasks and pays an observed $45–55/hr.
Quality Analyst
Review contributor work against rubrics for AI training and evaluation projects, document errors precisely, and feed recurring issues back into the guidelines. Contract, fully remote and async, with observed rates of $20–50/hr depending on domain background and project.
Excel Expert — Finance
Build and evaluate realistic Excel tasks — three-statement models, LBO/DCF, forecasting — that teach frontier AI models how finance professionals actually use spreadsheets. W-2 engagement through Cincinnatus LLC, US-based, 40 hrs/week weekdays through end of September and 20 hrs/week after.
ML Challenge Task Auditor
Audit applied machine-learning tasks — experiment designs, model-selection arguments, evaluation setups — that a frontier lab uses to train and test its models, and write rubric-anchored feedback on where the methodology holds and where it breaks. Observed pay runs $70–90/hr on Mercor, contract and remote.
Energetic Materials Expert for Redteaming
Mercor is recruiting explosives and energetic materials specialists to write adversarial prompts and grade frontier model responses against a misuse policy standard. Work is single-turn, writing-heavy, and paid per task in an observed $65–75 range.
Network Engineer - Data for Autonomous Systems annotation
Label and structure real production network data — logs, configs, telemetry, packet captures — so ML models can learn how networks fail and recover. The client wants working Tier 3 escalation engineers, not architects or SOC analysts, at 30–40 hours a week on Pacific time.
Drug Discovery Scientist
You design self-contained drug discovery research problems that frontier AI models must reason through — a prompt, a data room of real source-grounded evidence, and a grading document that scores the reasoning in a submitted memo. The answer must be derivable from the evidence you assemble but never stated in it, and you own the rework when a reviewer pushes back.
GPU Kernel Expert
Review GPU and accelerator kernel tasks — CUDA, Triton, NKI, Pallas — and judge whether they are numerically correct, fairly benchmarked, and properly scoped before they enter a frontier lab's training and evaluation data. Feedback is written against a rubric, roughly 40 hours a week, fully remote within the US.
Consulting Expert, Mandarin Chinese (mainland China) - AI Training
Create realistic consulting tasks, workspaces, prompts and grading rubrics for training AI agents in a mainland China business context. The role is for experienced management or strategy consultants who are native or fluent in Mandarin Chinese and have worked on market entry, operations and growth strategy in China.
Pharmacist / Pharmacy Technician — Medication Pronunciation Reviewer
Listen to short audio clips of spoken medication names and score whether each is pronounced accurately and intelligibly against a structured rubric. Work is batched and asynchronous, with an initial calibration exercise to align your scoring with the project standard; pay has been observed at $70/hr.
Systems & Integration Engineering Experts
Mercor is recruiting hands-on systems, integration, reliability, and test engineers to review and author hardware-grounded engineering problems, test procedures, failure analyses, and reference solutions used to train and evaluate AI models. The work is remote and project-based, with observed rates of $100–120/hr, and expects engineers who have personally owned bring-up, V&V, qualification, or production test on real physical products.
Legacy Codebase Migration Expert
Mercor is recruiting senior engineers who have personally led production migrations — Java 8 to 17, COBOL to Java, Oracle to Postgres, VMs to Kubernetes — to build and grade migration tasks for a frontier AI lab. Work is remote and async, with observed rates around $200/hr and the possibility of ongoing engagement as the program expands.
Partner & General Counsel - Legal Domain Expert
A senior practising lawyer works inside a frontier AI lab's research workflows, writing instruction specs and golden solutions for legal tasks and auditing model outputs that sound convincing but would not survive professional scrutiny. This is a full-time W-2 role through Cincinnatus LLC, hybrid and on-site in the Bay Area, with observed rates of $100–150/hr.
CVE Vulnerability Expert
Review CVE reproduction and remediation tasks built to train a frontier lab's models, judging whether the exploit actually fires, the fix actually closes it, and the Docker environment actually recreates the vulnerable conditions. Output is rubric-scored written feedback on each task, at roughly full-time hours, remote and US-based.
Business & Sales Operations Domain Expert
A full-time, hybrid Bay Area role writing instruction specs, golden solutions, and benchmarks that teach a frontier model how real revenue and sales operations work gets done. You are hired as a W-2 employee of Cincinnatus LLC and embedded with a leading AI lab's research team, working inside their tools and on-site several days a week.
Inorganic Chemistry & Synthesis Expert
Evaluate AI-generated reasoning about inorganic chemistry, materials synthesis, and experimental characterisation. The work includes checking synthesis pathways, diagnosing technical errors, and writing rigorous reference answers.
Biochemistry & Life Sciences Domain Expert
A full-time W-2 role with Cincinnatus LLC, placed inside a frontier AI lab's team to write instruction specs, golden solutions, and benchmarks for life sciences reasoning tasks. It is hybrid in the Bay Area with required on-site days, not a remote or project-based engagement, and observed rates run $65–105/hr.
Applied Physics Benchmark Specialist
Write and review expert-level physics multiple-choice questions — one correct answer, nine subtly wrong distractors, and a full chain-of-thought solution — for use as AI benchmark items. Work is asynchronous and remote at roughly 10+ hours per week, with observed pay of $61–77/hr.
Private Equity & Venture Capital - Finance Domain Expert
A leading AI lab is hiring a senior private equity or venture capital specialist to write instruction specs, produce golden solutions, and build benchmarks that test how frontier models reason about deal work. This is a full-time W-2 role through Cincinnatus LLC, hybrid on-site in the Bay Area, with an observed band of $110–150/hr for an initial six-month engagement.
Application Users - Linux Ubuntu 24.04 Expert - Professionals
Mercor is recruiting people who already run Ubuntu 24.04 as their everyday desktop to record real workflows in GNOME, Files, Settings, the terminal and apt for AI training data. You supply your own machine and a display that captures at over 2.5 megapixels; no licences, VMs or hardware are provided.
Human Resources Expert, Mandarin Chinese (mainland China) - AI Training
Create Mandarin Chinese prompts, workspaces, files and grading rubrics based on HR practice in mainland China. The work draws on at least five years of experience in HR business partnering, employee relations, compensation or talent.
STEM Expert, German (Germany) - AI Training
Create German-language STEM tasks, workspaces, files and grading rubrics that reflect professional scientific and engineering work in Germany. The role is remote and suited to scientists or engineers with at least five years of research or industry experience.
Investment Banking & M&A - Finance Domain Expert
A full-time W-2 role (employer of record: Cincinnatus LLC) embedded with a frontier AI lab's research team, writing instruction specs, golden solutions, and benchmarks for investment banking and M&A tasks. It is hybrid in the Bay Area — on-site multiple days a week, with relocation at your own cost if you are not already local.
General Physician
Evaluate health AI outputs for diagnostic accuracy, clinical reasoning, care quality, and patient safety. This six-week project is for currently practicing US physicians who can contribute at least 20 hours per week.
Software Developers (Must own MacBook)
Complete and narrate software development workflows on your own Mac to help evaluate and improve AI models. The work includes building, debugging, refactoring and reviewing code for approximately 10-20 hours per week.
Machine Learning Expert
Mercor keeps a standing pool of machine learning practitioners who write modeling problems, grade model-written training and evaluation code, and flag approaches that would train cleanly but fail in production. Comparable ML projects have posted at $70–120/hr, set per project; this listing is a pool entry, not a specific opening, so there is no accept/reject decision.
Film & Video Craft Expert (Camera, VFX, Editing & Performance)
Compare AI-generated videos and judge camera work, lighting, VFX realism, continuity, editing and performance against the original prompt. Write concise, production-specific rationales and participate in calibration with other film, theater and visual effects professionals.
Senior Counsel — Legal Domain Expert
A full-time W-2 role (via Cincinnatus LLC, discovered through Mercor) writing instruction specs, golden solutions, and benchmarks that teach a frontier AI model how real legal work is reasoned through. This listing is scoped to counsel and senior associate level — roughly 8 to 15 years post-qualification — and requires living in the Bay Area with on-site days at the client's office.
Preclinical R&D Expert – ADC & Bispecifics
Mercor is recruiting industry preclinical scientists with hands-on ADC or bispecific/multispecific antibody experience to annotate R&D data and model outputs for an AI lab building foundation models for drug discovery. Work is remote and hourly, with direct contact with the client team and observed rates of $60–100/hr.
Application Users - Origin on Windows - STEM
Mercor is recruiting scientists and engineers who use Origin or OriginPro on Windows in their actual research work to produce recorded application tasks and evaluate AI performance on them. You must already own a working Origin installation on a Windows PC and a display setup that captures screenshots above 2.5 megapixels — no licenses, VMs, or hardware are supplied.
LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness)
Mercor is staffing ML researchers to run scoped empirical experiments across vision and language — training image classifiers and generative models from scratch, fine-tuning open-weight LLMs, and hardening models against adversarial inputs and adversarial conversations. Work is project-based and remote, with observed rates of $100–120/hr depending on depth in a specialty area.
Payer & Market Access Expert — Gross-to-Net & Formulary Strategy
Mercor is staffing a short paid pilot in which US payer and market access professionals build and critique rubrics used to evaluate commercial drugs and drug-development programs. The core judgment call is whether an AI model's gross-to-net, rebate, and formulary-tier assumptions match how payers actually behave by drug class.
Stanford Youth Study — US Parents & Kids (Age 9–13)
Stanford University and Bocconi University are enrolling US parent–child pairs (child aged 9–13) in a study on how smartphones affect family life and youth well-being. Participation involves roughly an hour of online surveys, a few text-message check-ins, and installing a parental-control app on the child's phone, with an observed flat payment of $120 on completion.
Molecular Biology Experts - US
Design nucleic acid constructs — primers, plasmids, gRNAs, mRNA and HDR templates — as ground-truth training tasks for a frontier AI lab's molecular biology models. W-2 employment through Cincinnatus LLC, minimum 20 weekday hours per week, observed pay $70–105/hr.
Chemical Defense Researcher
Mercor is recruiting chemical defense practitioners to red-team frontier models on the dual-use line — writing benign, dual-use, and adversarial prompts, then judging whether responses met a defined policy standard. Observed pay is $65–75 per task, and every judgment requires a written technical rationale a non-specialist can follow.
Sales and Marketing Expert
Mercor is contracting senior enterprise sales, marketing, and RevOps operators to write evaluation tasks and grading rubrics for AI systems working in Fortune 500 go-to-market contexts. You build realistic scenarios — $1M+ ACV deals, buying committees, brand campaigns — draft reference answers, and define what separates real operator judgment from playbook recitation.
PDF Annotation & Transcription Experts – Malayalam
Annotate the structure of real Malayalam PDF pages — bounding and typing every region, setting reading order, and transcribing all text character-for-character in Malayalam script. Every task is reviewed end to end by a second Malayalam expert, and experienced annotators take on that review work.
Voice Actor: CX Agent Voice Cloning (Standard French)
Record standard, neutral French voice samples that will be used to train and clone a voice for a client's internal customer-experience AI agent. Requires professional voice work experience, a near-professional home recording setup, and explicit consent to voice cloning within this single use case.
Voice Actor: CX Agent Voice Cloning (Spanish - Mexico)
Record scripted Mexican Spanish voice sessions that will be used to train and clone a voice for a client's internal customer-experience AI agent. Requires a native Mexican Spanish accent, residence in Mexico, a near-professional home recording setup, and explicit consent to voice cloning for this single use case.
Voice Actor: CX Agent Voice Cloning (Canadian French)
Record scripted and conversational Canadian French voice samples that will be used to train and clone a voice for a client's internal customer-experience AI agent. Requires native Canadian French fluency with neutral international French delivery, a near-professional home recording setup, and explicit comfort with voice cloning.
Voice Actor: CX Agent Voice Cloning (Spanish - Peninsular)
Record Castilian Spanish voice samples that will be used to train and clone a customer-experience AI agent's voice. Requires Spain-based native Peninsular speakers with professional voice work experience and a near-studio home recording setup.
PDF Annotation & Transcription Experts – Telugu
Annotate the structure of real Telugu PDF pages — bounding and typing every region, setting reading order, and transcribing all text character-for-character in Telugu script. Work is remote and task-based at an observed $10–11/hr, with every task reviewed end to end by a second Telugu expert.
Accounting & Audit Domain Expert
A full-time W-2 role through Cincinnatus LLC, placed inside a frontier AI lab's team to write instruction specs, golden solutions, and benchmarks for accounting and audit reasoning. It is hybrid in the Bay Area — on-site multiple days a week when required, with no relocation assistance.
Trainium (NKI) Kernel Expert
Review NKI kernel work — CUDA→NKI migrations, Trainium performance optimizations, and numerical-correctness claims — and write rubric-scored feedback that a frontier lab uses as training and evaluation data. Full-time remote engagement at roughly 40 hours per week, US-based, observed at $70–90/hour.
Radiologicals Safety for Redteaming
Mercor is recruiting radiological safety practitioners to write adversarial prompts and grade frontier-model responses on the dual-use line between legitimate radiation work and misuse. Observed pay is $65–75 per task, remote and asynchronous, with every judgment requiring a written technical rationale.
Agent Engineer
Mercor is recruiting engineers who have shipped LLM agents that real users depended on, and who can speak in detail about reliability, evaluation, and internal adoption. The process is a conversational AI screen followed by a paid 30-minute live conversation, observed at $100–500 depending on depth of experience.
German Language and Literature Expert (German)
Evaluate and rewrite German text produced by a large language model — translation, summarization, dialogue and grammar correction — and annotate it for grammatical, syntactic, semantic and stylistic features. Mercor lists this at an observed $50/hr, remote and largely asynchronous, for people with professional editing, criticism, or linguistic annotation experience in German.
Disease-Area Clinician — Trial Endpoints & Prescribing
Mercor is recruiting US-based, board-certified practising physicians to build and critique rubrics that AI systems use to analyze commercial drugs and drug-development programs. The work centers on reading trial endpoints and judging, in your own therapeutic area, whether a result would actually change how you prescribe.
Customer Success Engineer (LatAm)
Front-line technical support for Mercor's talent platform, investigating bug reports across an AI + SaaS stack and deciding what is a real engineering defect versus UX friction or a model edge case. Remote from Latin America, six days a week, with substantial overlap with Pacific Time.
Legal Expert, Mandarin Chinese (mainland China) - AI Training
Experienced lawyers will create realistic mainland China legal tasks, prompts, files and grading rubrics for training AI agents. The role requires at least five years of practice, native or fluent Mandarin Chinese, specified PRC qualifications and availability for about 21 hours per week.
Propulsion & Effects Engineer
Mercor is recruiting propulsion, ordnance and effects engineers to write and grade adversarial prompts that test whether frontier models can separate legitimate technical questions from misuse attempts. Work is single-turn prompt authoring plus written adjudication against a policy standard, observed at $65–75 per task.
Audiobook QA Expert — Dutch
Listen to AI-narrated Dutch (Netherlands) audiobooks and log exactly where the synthetic narration breaks — mispronunciations, skipped or added words, misread numbers and abbreviations, flat or wrong intonation, audio artifacts. Roughly 20 hours per week, remote, with observed pay of $15–20/hr.
Application Users - Unreal Engine on Windows - Generalist
Mercor is recruiting Unreal Engine practitioners to work inside the editor on their own Windows machines, producing screenshot-backed records of real workflows for AI training and evaluation. You must already have Unreal installed and be able to open it today — no licences, VMs, or hardware are provided, and your display must support screen captures above 2.5 megapixels.
Expert Interviewer
Run ~20-minute structured interviews with shortlisted engineers in GPU kernel development, security research, and ML-compiler/accelerator work, then tier each one against a provided rubric. Roughly 10 hours a week, fully remote, scheduled inside weekday business hours.
News Analysts, Reporters, and Journalists
Mercor is contracting working journalists to evaluate AI-generated news copy, scripts, and analyses for accuracy, sourcing rigor, and editorial soundness. It is a task-based engagement of roughly four to six weeks at 20+ hours per week, with observed rates in the $70–110/hr band once you clear the first approved task.
Biotechnology Expert
Mercor keeps a standing biotechnology pool that frontier AI labs draw from when they need process, assay, and developability judgment their models can't produce. You write development problems with graded reference answers and score model attempts on whether the science would actually survive scale-up, release testing, and regulatory review.
Fitness Data Contributor — $60
Mercor pays a one-time $60 for a single de-identified Apple Health export that combines your on-device HealthKit data with clinical records pulled from a qualifying US health system. It is a data contribution, not an evaluation task: you need an iPhone, a rich HealthKit history, and a linked EHR that includes clinical notes.
Cloud / DevOps Engineer (Infra & IaC)
Write and evaluate infrastructure engineering tasks — Kubernetes cluster failure diagnosis, Terraform/CDK design, AWS service integration, CI/CD reasoning — used to train and assess frontier language models. Full-time (40 hrs/week) W-2 engagement through Cincinnatus LLC with placement on a leading AI lab's GenAI team; observed pay band $75–110/hr.
Medical Expert, Mandarin Chinese (mainland China) - AI Training
Physicians will create realistic medical tasks, prompts, workspaces and grading rubrics grounded in clinical practice in mainland China. The remote pilot calls for about 21 hours per week and evaluates both medical expertise and the ability to judge AI outputs consistently.
Biology Expert
Mercor keeps a standing biology pool that frontier AI labs draw from when they need experimental judgment their models can't produce. You write research problems with reference answers, grade model-written designs and interpretations, and flag answers that are textbook-correct but would fail at the bench.
Mechanical & CAD Engineers (Must own MacBook)
Complete and narrate mechanical design and CAD workflows that will be used to benchmark and improve AI models. You must have relevant engineering training and own a Mac with an Apple M-series chip or Intel processor running macOS 15 or higher.
Inorganic Materials, Semiconductor & Superconductor Experts
Mercor is hiring experimental materials scientists — inorganic synthesis, superconductors, semiconductors and advanced packaging — to build and grade the scientific data a frontier lab uses to train physical-science models. Work is remote, US-based, and runs from 10 to 40 hours per week at an observed rate near $84/hr.
Legal Expert, German (Germany) - AI Training
Create realistic German legal tasks, prompts, files and grading rubrics used to train and evaluate AI agents. This remote pilot calls for an experienced Volljurist with native or fluent German and substantial knowledge of German civil, commercial, corporate and procedural law.
Program Lead
Lead day-to-day delivery for AI training data, human data and model evaluation programs, coordinating large pools of domain experts against quality and throughput targets. The role combines contributor operations, guideline development, submission evaluation, metric tracking and cross-functional issue resolution.
Medicine Expert (M.D./D.O.)
Mercor keeps a standing pool of board-certified physicians who write clinical vignettes and grade AI model output on diagnostic reasoning, workup, and management. Joining is an application plus a roughly 20-minute AI interview and license verification; paid work comes later, when a client project matches your specialty.
Central Bank Expert — Non-G7 Policy Communication
Mercor is staffing country-specific experts to interpret non-G7 central bank communication — statements, minutes and speeches in the source language — and score them on a dovish-to-hawkish scale from a fixed evidence cutoff. You also grade model-generated macro analyses against your own standard, with observed rates of $150–250/hr.
Audiobook QA Expert — French
Listen to AI-narrated audiobooks in French (France) and log every place the synthesized voice slips — mispronunciations, skipped or added words, misread numbers, flat intonation, audio artifacts. Roughly 20 hours a week, remote, with issues marked to the exact span in the platform's review tooling.
Medical Expert
Mercor keeps a standing pool of credentialed clinicians who write clinical cases, grade model reasoning, and mark where textbook-correct advice fails a real patient. Joining the pool is not a job offer — you are invited to specific projects, at rates that have posted between $60 and $180 an hour depending on profession and scope.
Infrastructure / Site Reliability Engineer (SRE)
Build and operate enterprise-grade production infrastructure across Kubernetes and AWS, with responsibility for reliability, observability, deployments, and incident response. This full-time engagement suits engineers with hands-on experience debugging and scaling complex distributed systems.
Audiobook QA Expert — English (US)
Listen to AI-narrated audiobooks in US English and log every place the synthesized voice slips — mispronunciations, dropped or added words, misread numbers, wrong abbreviation expansions, flat or misplaced intonation, and audio artifacts. Roughly 20 hours a week, fully remote, with each issue categorized and timestamped to the exact span in the review tool.
Generalist (Macbook User)
Mercor is staffing a research project with a frontier AI lab that requires contributors to work on an Apple Silicon MacBook (M-series, macOS 15+). You'll write research questions, model answers, and evaluation materials used to benchmark and improve language models, roughly 10–20 hours per week at an observed $50–60/hr.
Clinical Medicine Domain Expert
A practicing physician role embedded with a frontier AI lab's research team, writing instruction specs, golden solutions, and clinical benchmarks that define what a correct medical answer looks like. Full-time W-2 employment through Cincinnatus LLC, hybrid on-site in the Bay Area for an initial six-month engagement.
Voice Actor: CX Agent Voice Cloning - Northern UK
Record scripted and conversational voice samples in an authentic Northern English accent for a client's internal customer-experience AI agent, with the explicit understanding that your voice will be cloned for that single use. Work is remote and self-recorded from your own treated space, at roughly 5–10 hours per week; Mercor has listed this at $50/hr.
Voice Actor: CX Agent Voice Cloning (Australian English)
Record scripted and conversational voice samples in an authentic Australian accent for a customer-experience AI agent built by a research lab Mercor partners with. Your voice may be cloned for that single internal product — the listing states recordings will not be sold, licensed, or reused elsewhere.
Machine Learning Expert
Develop difficult machine learning engineering task ideas for a long-horizon research project, with an emphasis on inference, evaluation and systems problems. You must currently or recently have worked in an ML research role and have personally trained, fine-tuned or optimized models using PyTorch or JAX.
Legal Expert
Mercor keeps a standing legal pool that frontier AI labs draw from when they need practitioner judgment to build and grade evaluation sets. You write legal problems from real matters, produce the reference answers models are graded against, and mark up model-drafted work where the law is stated correctly but the argument would still lose.
Senior Character Animator
Review AI-generated character performances against source acting, then identify failures in timing, body mechanics, facial nuance and emotional intent. This part-time W-2 role requires senior production credits and periodic in-person working sessions in Los Angeles or New York.
Biologist Talent Network
Mercor is building a standing pool of biologists to write, review, and grade tasks that stress-test frontier AI models on molecular and cellular biology, experimental design, and life-sciences data analysis. This is an open application rather than a single project — you complete a profile, pass an AI interview, and get matched to contracts as labs request biology coverage.
Chemist Talent Network
Mercor maintains a standing network of chemists who are matched to AI training and evaluation projects as labs request them. You apply once, complete an AI interview, and are contacted when a project needs your specific analytical or synthetic background.
Software Engineer (AI-Native Platform & Integrations)
Build and ship a multi-tenant AI-native SaaS platform end-to-end — integrations layer, real-time analytics, and an MCP-style agentic tool runtime — through pilot launch with real users. Hired via Mercor's talent network as a W-2 role administered by Cincinnatus, with observed rates of $70–110/hr.
Marketing Specialist Talent Network
Mercor is building a standing pool of marketing practitioners to evaluate and train AI models on campaign, research, and brand work. You apply once, pass an AI screening interview, and get matched to contract projects as labs request marketing expertise — observed rates for this network run $60–80/hr.
Mathematician Talent Network
Mercor maintains a standing pool of mathematicians who are matched to AI lab projects as they open — writing problems, grading model proofs, and flagging where reasoning breaks down. Applying is a one-time profile plus an AI interview; work arrives on a rolling basis at an observed $60–80/hr, typically 15–30 hours per week.
Management Consultant Talent Network
Mercor is assembling a bench of experienced strategy consultants to grade AI-generated consulting deliverables — market analyses, models, decks, and implementation plans — against task-specific rubrics. This is a talent-network application rather than an open project; observed rates for comparable Mercor expert work sit in the $100–150/hr range when engagements activate.
Nursing Talent Network
Mercor maintains a standing pool of licensed nurses who are matched to AI training and evaluation projects as labs request clinical expertise. You apply once, complete an AI-led screening interview, and get contacted when a project fits your background — typical engagements run 15–30 hours per week.
Full-Stack Engineer Talent Network
Mercor maintains a standing pool of full-stack engineers who write, review, and grade tasks used to train and evaluate frontier coding models. You apply once, pass an AI screening interview, and get matched to projects as labs request full-stack expertise — typically 15–30 hours per week at observed rates of $70–150/hr.
DevOps / Platform Engineer Talent Network
Mercor maintains a standing pool of DevOps and platform engineers who write, review, and grade infrastructure tasks used to train and evaluate frontier AI models. You apply once, pass an AI-led screening interview, and get matched to projects as labs request DevOps expertise — typically 15–30 hours per week at observed rates of $70–150/hr.
Physician Talent Network
Mercor maintains a standing pool of licensed physicians who are matched to AI lab projects as they open, rather than to one fixed role. You apply once, complete an AI-led screening interview, and get contacted when a project needs your specialty — typically 15–30 hours per week at observed rates of $110–250/hr.
Physicist Talent Network
Mercor maintains a standing network of physicists who are matched to AI training and evaluation projects as labs request them. You apply once, pass an AI screening interview, and get contacted when work matching your subfield opens up — typically 15–30 hours per week at an observed $60–80/hr.
Management Consultant Talent Network
Mercor maintains a standing pool of management consultants who build and grade AI training tasks — strategy cases, financial and operating models, client-facing deliverables. You apply once, pass an AI screening interview, and get matched to project work as labs request consulting expertise, typically at 15–30 hours per week.
Machine Learning Engineer Talent Network
Mercor maintains a standing pool of machine learning engineers who are matched to AI lab projects as they open, rather than to one fixed role. Work involves building realistic ML engineering tasks, grading model outputs against them, and writing feedback that explains where a model's reasoning or code breaks down.
Backend Engineer Talent Network
Mercor maintains a standing pool of backend engineers who write, review, and grade tasks used to train and evaluate frontier AI models on API design, databases, and distributed systems. You apply once, complete an AI screening interview, and get matched to contract projects as labs request backend expertise — typically 15–30 hours per week at observed rates of $70–150/hr.
SOC Investigation Specialist Talent Network
Mercor is recruiting Tier 2+ SOC analysts to evaluate AI-generated security investigations and produce ground-truth cases across SIEM, endpoint, cloud, and identity data. Work is Splunk-heavy: you pivot across logs, reconstruct timelines, and make accept/reject calls on whether an investigation's conclusions hold up.
HR & Administration Specialist Talent Network
Mercor is building a standing pool of HR and administration practitioners who evaluate and train AI models on recruiting, payroll, benefits, and compliance work. You apply once, complete an AI screening interview, and get matched to contract projects as labs request HR expertise — observed rates for this network run $60–80/hr.
Data Scientist Talent Network
Mercor is pre-screening experienced data scientists for future AI evaluation projects with frontier research labs. There is no open project today — accepted applicants sit in a talent pool and are contacted when matching work appears.
Compliance & Risk Specialist Talent Network
Mercor is building a standing pool of compliance and risk professionals to write, review, and grade AI model outputs on regulatory frameworks, controls testing, and risk assessment. This is an open application rather than a single project — you complete a profile, pass an AI screening interview, then get matched to labs' projects as they open, typically 15–30 hours per week at an observed $60–80/hr.
Financial Analyst Talent Network
Mercor maintains a standing pool of financial analysts who build and grade tasks used to train and evaluate frontier AI models on modeling, forecasting, and statement analysis. You apply once, pass an AI interview, and get matched to contract projects as labs request finance expertise — typically 15–30 hours per week at observed rates of $60–180/hr.
Graphic & UX/UI Designer Talent Network
Mercor maintains a standing pool of graphic and UX/UI designers who evaluate AI model output, write design tasks drawn from real client work, and give written critique that feeds model training. This is an open application rather than a single project — you pass an AI screen once, then get matched to contracts as labs request design expertise.
Physical Scientist Talent Network
Mercor maintains a standing pool of physicists, chemists, materials scientists, and other physical scientists who write and grade tasks used to train and evaluate frontier AI models. You apply once, pass an AI screening interview, and get matched to contract projects as labs request physical sciences coverage — typically 15–30 hours per week at an observed $60–80/hr.
Business Intelligence Analyst Talent Network
Mercor is building a standing pool of business intelligence practitioners to write, review, and grade tasks that test how well frontier models handle SQL, dashboarding, and data modeling work. You apply once, pass an AI screening interview, and get matched to contract projects as labs request BI expertise — typically 15–30 hours per week at an observed $70–120/hr.
Frontend Engineer Talent Network
Mercor maintains a standing network of frontend engineers who write, review, and grade tasks used to train and evaluate AI coding models. You apply once, pass an AI-led screen, and get matched to project work as labs request frontend expertise — typically 15–30 hours per week, remote and async.
Lawyer Talent Network
Mercor maintains a standing network of practicing and former lawyers who write, review, and grade legal tasks used to train and evaluate frontier AI models. You apply once, pass an AI screening interview, and get matched to contract projects as labs request legal expertise — typically 15–30 hours per week at observed rates of $60–150/hr.
Open Source Applied Engineer Talent Network
Mercor is building a bench of active open source contributors to design coding evaluations and test cases across Python, Java, C, JavaScript, and TypeScript. Work is remote and asynchronous, with observed rates of $100–120/hr depending on language depth and contribution history.