The work
You will create realistic computational earth-science problems, write precise task instructions, build reproducible environments, and prepare reference solutions and evaluation criteria. Day to day, this may include inspecting model-generated Python, geospatial analyses, simulations, forecasts, maps, or other technical deliverables for substantive scientific and computational errors.
What the screening covers
The AI interview is likely to verify prior paid AI-training or model-evaluation work, practical Python proficiency, and depth in at least one relevant specialization. Expect follow-up questions about coordinate reference systems, spatial and temporal data, numerical methods, model assumptions, uncertainty, reproducibility, and validation against observations or established models. Experience creating rubrics, graders, benchmark tasks, or reference answers is preferred.
Evidence of expertise
Strong applications include public evidence such as peer-reviewed research, preprints, research profiles, open-source contributions, analysis code, datasets, models, notebooks, maps, simulations, or documented forecasting and remote-sensing projects. The strongest evidence makes your individual contribution and use of computational methods easy to verify.
Logistics
This is a global, fully remote contractor role with a flexible schedule and a stated commitment of at least 15 hours per week. The listing does not state that all collaboration is asynchronous, so candidates should allow for the AI interview, hiring-manager review, onboarding, and any project coordination. Work is paid per accepted task, and task availability and completion time may vary.
Pay band: $60–120/hr