Computational Materials Science Expert

$150–200/hrrate as listed

AfterQuery is contracting published computational materials researchers to build and evaluate ML tasks on real atomistic, electronic, and structural datasets. Work is remote and asynchronous over a 2–3 week window, with an observed rate of $150–200/hour at 10–20 hours per week on projects you accept.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work looks like

You take on discrete, project-scoped tasks that sit between machine learning and computational materials science. In practice that means writing and evaluating ML pipelines against simulation-derived data — DFT relaxations and band structures, MD trajectories, Monte Carlo sampling output — and producing property-prediction, materials-discovery, or structure-property modelling artefacts with the reasoning made explicit. Tasks frequently ask you to build a defensible baseline, state your assumptions, and document why a given featurisation, target, or validation split is or isn't appropriate for the underlying physics. Some tasks are generative (write the reference solution); some are evaluative (judge whether a model's output or a chain of reasoning holds up against what the simulation data can actually support).

What the screen is actually checking

The first-author publication requirement is a real gate and is verified — the platform wants people who have defended methodology in review, not people who have read about it. Beyond that, screening probes whether your materials knowledge and your ML knowledge are genuinely joined: whether you know why random k-fold on a materials-project-style dataset leaks, why an interatomic potential fit on near-equilibrium configurations fails in the repulsive regime, what a DFT functional choice does to the label noise on your target property. Expect follow-ups that go one layer deeper than your first answer. Vague fluency in both fields separately reads as a fail; specific, falsifiable statements about a system you actually computed read as a pass.

Logistics

Fully remote, fully asynchronous — no standing meetings, no timezone requirement.
10–20 hours per week on accepted projects, over a 2–3 week engagement window.
Project-based; you accept or decline individual assignments.
Pay observed in the $150–200/hour range, not guaranteed and generally tied to task complexity and calibration performance.
Deliverables are code plus written methodology, so reproducibility and clear documentation matter as much as model performance.

What they look for

First-author peer-reviewed publication in a computational materials, chemistry, or physics field
Hands-on ML model development in Python (PyTorch, scikit-learn, or equivalent) applied to scientific data
Working experience with DFT, molecular dynamics, or Monte Carlo simulation codes (VASP, Quantum ESPRESSO, LAMMPS, ASE)
Master's or PhD in materials science, computational chemistry, physics, CS, or a related quantitative field
Familiarity with materials-specific ML architectures (CGCNN, M3GNet, MatGL, ALIGNN, or comparable GNNs)
Clear technical writing: documenting assumptions, validation choices, and failure modes reproducibly

Included with this listing

Interview prep — 10 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether the publication credential is real and whether you personally did the computational work rather than supervising or co-authoring around it.
A strong answer: Name the journal and the system studied, then get concrete fast: the code used, the functional or force field, the convergence parameters you settled on, and the one methodological decision a reviewer pushed back on.
Where people lose it: Describing the paper's conclusions at abstract level without ever naming a code, a basis set, a cutoff, or a system size — that pattern reads as someone who wrote the discussion section while a collaborator ran the calculations.