STEM Expert AI Evaluation

About Open Train

Open Train is the #1 platform for finding and building careers in AI training and data labeling. Open Train AI is hiring contractors to help improve advanced AI systems through expert evaluation, scientific reasoning, and technical problem-solving.

• Fully remote work from an eligible country
• Flexible schedule with approximately 15 hours per week
• Contractor and part-time engagement
• Advertised pay range of $50-$100 per hour

AI training is the human side of building artificial intelligence. Experts review model responses, solve challenging problems, and provide accurate examples so AI systems can reason more reliably across technical and scientific subjects.

• Evaluate AI-generated answers for accuracy and sound reasoning
• Create reference solutions that help models learn
• Work on text generation and question-answering tasks
• Apply specialized knowledge to cutting-edge AI development

As a STEM Expert, you will contribute your deep knowledge in at least one scientific or technical discipline. Your work may involve technical problem-solving, scientific reasoning, data analysis, Python-based calculations or simulations, and careful review of AI outputs.

Candidates do not need expertise across every STEM discipline. Relevant expertise may come from academic research, industry, teaching, engineering, independent technical work, or other demonstrated experience. A specific degree or number of years of experience is not strictly required when strong expertise can be demonstrated.

• Pay: $50-$100 per hour
• Engagement:
  Contractor, part time
• Expected commitment:
  Under 20 hours per week, approximately 15 hours
• Project duration:
  Expected to be under one month
• Location:
  
  Remote from listed eligible countries
• Working language:
  English
• Availability:
  Candidates should be ready to begin promptly

You will solve, review, validate, and explain challenging scientific, mathematical, and technical problems within your area of expertise. Depending on the task, you may interpret equations, datasets, experimental results, technical diagrams, or scientific literature.

• Solve complex scientific, mathematical, or technical problems
• Review AI-generated solutions for factual, mathematical, and scientific accuracy
• Find incorrect assumptions, reasoning errors, missing constraints, and subtle technical mistakes
• Develop clear, accurate, and reproducible reference solutions
• Use Python for scientific computation, data analysis, simulations, or validation
• Perform sanity checks on numerical and analytical results
• Compare alternative approaches and assess whether conclusions are evidence-based
• Explain technical reasoning, findings, and corrections clearly

You should have strong expertise in at least one relevant STEM or scientific domain and practical proficiency with Python. Experience with tools such as Num Py, Sci Py, pandas, Sym Py, Jupyter, or other scientific libraries may be useful, but no specific Python library is mandatory.

The work requires careful quantitative reasoning and the ability to identify assumptions, constraints, edge cases, and possible sources of error. Experience validating, reviewing, analyzing, or troubleshooting technical work is important.

• Deep expertise in at least one relevant STEM or scientific field
• Practical proficiency with Python
• Strong scientific, mathematical, analytical, or quantitative problem-solving ability
• Ability to explain complex technical concepts clearly
• Experience with technical review, validation, analysis, or troubleshooting
• Experience with text generation and question-answering tasks
• Fluent English proficiency
• Expertise in STEM scientific reasoning and AI evaluation

Candidates may qualify through strong expertise in one or more of the following areas. Adjacent scientific, engineering, or quantitative disciplines may also be considered when directly relevant.

• Chemistry
• Biology
• Mathematics
• Physics
• Earth science, geoscience, or environmental science
• Linguistics or computational linguistics
• Robotics
• Other directly relevant scientific, engineering, or quantitative disciplines

Compensation is output-based within the advertised $50-$100 per hour range. Payment is based on completed tasks that meet project specifications, and the time required for each task may vary based on complexity, experience, and workflow.

Minimum submission requirements may apply. If selected, you may be expected to begin your first task within 24-48 hours after completing onboarding.

• Flexible working days and hours, including weekends if desired
• Output-based compensation
• Task completion standards apply
• Prompt start expected after onboarding

Applicants must be based in one of the listed countries. This project is not open worldwide.

• Australia, Austria, Belgium, Brazil, Canada, Chile, Colombia, Denmark, Egypt, Finland, France, Georgia, Germany, Greece, Iceland, India, Italy, Japan, Mexico, Monaco, Morocco
• Netherlands, New…