Chemistry Reasoning Evaluator — BS/MS/PhD in Chemistry (Top-100 Univ. preferred)

About OpenTrain

OpenTrain aggregates data-labeling and AI-training work from many companies and platforms into a single job board. Creating an OpenTrain account is free and applying takes only a few minutes.

We connect specialists and reviewers with short- and long-term projects that help AI systems learn from human examples. This role is posted through OpenTrain's network of AI evaluation projects.

About AI Training Work

AI training (also called data labeling, annotation, or human feedback) is the human side of building models: people evaluate model outputs, correct mistakes, and create exemplar responses that teach models how to reason and communicate.

As an evaluator you will directly shape how chemistry-focused AI systems judge correctness, explain reasoning, and handle safety-sensitive or technical claims.

The Role

Title: Chemistry Reasoning Evaluator — entry level (educational requirements apply).

Engagement: Contractor, part-time, remote, worldwide.

Compensation: Paid per hour at USD 80/hr.

Time requirement: Minimum availability of 17–20 hours per week (typically under 20 hrs/week); during active sprints a preferred cadence is around 8 hrs/day.

• Data type: Text — you will evaluate written AI responses.
• Labeling task: Evaluation/Rating using detailed rubrics.
• Platform/software: OTHER (project-specific tools).

What You'll Do

You will review AI-generated chemistry responses and assess them against detailed rubrics for correctness, depth of reasoning, clarity, and safety. Work is careful, methodical, and evidence-based.

Tasks include error identification, fact-checking, drafting exemplar solutions, and rating or ranking multiple model outputs to guide model improvements.

• Assess correctness of calculations, unit consistency, and dimensional analysis.
• Evaluate reaction mechanisms, stoichiometry, thermodynamics and kinetics reasoning.
• Check spectroscopy and analytical-method interpretations.
• Spot methodological or conceptual errors and unsafe assumptions.
• Draft clear, step-by-step exemplar explanations and model solutions.
• Compare and rate multiple responses using detailed rubrics.

Requirements

Must-have educational qualification and domain foundation are strict requirements for this role. Onboarding includes paid evaluation exercises to confirm alignment with project rubrics.

• BS, MS, or PhD in Chemistry or a closely related chemical science (Top-100 university preferred).
• Strong foundation across general, organic, inorganic, physical, and analytical chemistry.
• Good laboratory and safety literacy.
• Excellent scientific writing in clear, step-by-step English (C1+ level) with correct notation and units.
• Quantitative rigor: dimensional analysis, unit consistency, and awareness of approximations and uncertainty.
• Effective fact-checking using reputable public sources and precise, consistent referencing when required.
• Consistent application of evaluation rubrics, meticulous attention to detail, and reproducibility mindset.
• Availability: minimum 17–20 hrs/week; preferred cadence ~8 hrs/day during active sprints.
• Onboarding: a paid 1–2 hour qualification exam and a paid 1–2 hour project exam.

Preferred And Bonus Qualifications

The following are not required but will make applications more competitive and may be helpful on technical or specialized projects.

• Research experience, analytical writing, or debate experience.
• Programming literacy (e.g., Python or Matlab) and familiarity with LaTeX for clear equations.
• Prior data labeling, RLHF, or AI model-evaluation experience (bonus).

How It Works / How To Apply

Create a free OpenTrain account and submit your application. Applications are typically quick and require your educational background, availability, and a brief statement of relevant experience.

If shortlisted, you will complete a paid 1–2 hour qualification exam and a paid 1–2 hour project exam to demonstrate your chemistry reasoning and alignment with the rubrics.

• Employment type: Contractor, part-time — remote work from anywhere.
• Compensation: USD 80 per hour, paid per project terms.
• Tasks are text-based evaluation ratings; tools and exact workflows will be provided during onboarding.