What the work involves
You will spend most of your time writing and judging technical content rather than running tools. A typical block might mean authoring an expert-level problem on nucleation delay for a thermal ALD Al₂O₃ process on a hydrogen-terminated silicon surface, then producing a reference solution with the reasoning made explicit. Other blocks are evaluative: reading model-generated explanations of precursor decomposition pathways, ALD window boundaries, or GPC anomalies, marking where the chemistry is wrong or merely plausible-sounding, and writing a short rationale that a reviewer who is not in your subfield can follow. A third stream is structuring — taking process parameters, recipes, and characterization results and rendering them into clean, consistent, model-ready records.
- Draft and solve problems spanning ALD process development, precursor selection, surface reaction mechanisms, and thin-film integration
- Review AI-generated scientific reasoning for factual and mechanistic errors, not just tone or formatting
- Rate and rank competing outputs against defined scientific rubrics with written justification
- Normalize technical knowledge — temperatures, pulse/purge timings, saturation curves, characterization outputs — into structured formats
What the platform screens for
Mercor's screen is AI-led and interview-based, and it leans hard on specificity. Expect follow-ups that push past your résumé summary into what you actually did at the bench: which precursors, which reactor geometry, what failure mode you chased and how you diagnosed it. Vague process descriptions get probed until they either resolve into detail or don't. The screen also tests evaluation judgment — whether you can distinguish a wrong answer from an unconventional-but-correct one, and whether you can articulate why in writing. An advanced degree (PhD or MS) in materials science, chemistry, chemical engineering, or physics is the usual entry point, though equivalent hands-on industrial experience is considered.
Logistics
Remote and US-based, structured as a long-term ongoing engagement rather than a one-off project. Committed availability runs from a 10-hour weekly minimum up to 40 hours, and most work is asynchronous — tasks are pulled from a queue with defined turnaround windows rather than scheduled meetings. The $84/hr figure reflects rates observed on this listing; actual offers vary with credentials, screen performance, and task type, and are not guaranteed.