Help improve AI systems by evaluating chemical safety questions and model responses for accuracy, risk, and appropriate access. This remote contract offers $70-$100 per hour with a 20+ hour weekly commitment.
About OpenTrain
OpenTrain AI contracts specialists to help build safer, more capable artificial intelligence. OpenTrain is the platform where people discover AI training work, build a professional profile, and grow a lasting portfolio in this fast-growing field.
- Remote contract and part-time opportunity
- Build experience in specialized AI training and evaluation
- Create an OpenTrain account and apply in minutes
About AI Training Work
AI training is the human side of developing modern models. Experts review prompts and answers, identify errors and unsafe behavior, and provide examples that help AI systems respond more accurately and responsibly.
In this role, your chemical safety expertise will help shape how models handle sensitive technical questions, including requests with legitimate, dual-use, or adversarial purposes.
- Contribute directly to the development of advanced AI systems
- Use professional judgment to improve safety and reliability
- Work remotely with flexible part-time scheduling
The Chemical Safety AI Model Evaluator Role
OpenTrain is seeking a Chemical Safety AI Model Evaluator to assess how artificial intelligence systems respond to sensitive questions involving chemistry, toxicology, industrial hygiene, process safety, and chemical security.
The listing identifies this as an entry-level opportunity, but the work requires relevant professional expertise or a comparable subject-matter background. The expected commitment is 20 or more hours per week.
- Pay: $70-$100 per hour
- Payment type: Hourly contract
- Work arrangement: Remote
- Engagement: Contractor, part time
- Language: English
- Data type: Text
What You’ll Do
You will review realistic chemical safety requests and multiple AI-generated answers. Your evaluations will help determine whether model behavior is accurate, safe, appropriate, and suitable for the context.
You will also create detailed prompts and responses that support learning across diverse chemical safety topics. The work includes testing models for inaccuracies, bias, unsafe behavior, and inappropriate refusals.
- Classify requests as benign, dual-use, or adversarial
- Evaluate three model answers for safety, accuracy, and appropriateness
- Explain the reasoning behind each judgment
- Rank model responses from strongest to weakest
- Write the answer the model should have provided
- Identify why a dual-use request is sensitive
- Explain the legitimate purpose and appropriate access boundaries for sensitive information
- Create prompts and responses for chemical safety model training
Required Expertise and Skills
You should be able to distinguish benign, dual-use, and adversarial chemical questions using sound safety judgment. You must also be able to evaluate technical answers for accuracy, risk, context, and appropriate boundaries, then explain your decisions clearly in writing.
A relevant degree or comparable professional background is typically expected. Experience with safety data sheets, exposure assessment, incident reports, or regulatory review is valuable.
- Expertise in synthetic or analytical chemistry, toxicology, industrial hygiene, process safety, chemical security, emergency response, or chemical regulation
- Ability to assess safety, accuracy, and risk in technical chemical information
- Ability to explain model-response judgments and write an appropriate corrected answer
- Strong written communication and clear technical reasoning
- Familiarity with safety data sheets, exposure assessment, incident reports, or regulatory review
- Relevant experience connected with EPA, OSHA, HSE, OPCW, hazmat operations, poison control, or chemical incident response is helpful
Eligibility and Availability
The role description highlights availability to citizens or permanent residents of the United States, United Kingdom, Canada, Australia, and New Zealand. The structured opportunity listing also includes Austria, Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, and Sweden.
Confirm your eligibility and location during the application process. English is the required working language, and the expected workload is 20 or more hours per week.
- Eligible locations are specified in the opportunity listing
- English-language work
- 20+ hours per week
Build Your AI Training Career
OpenTrain helps specialists manage AI training opportunities, demonstrate credible work, and develop a professional portfolio. Your evaluations can give you practical experience at the intersection of chemical safety and cutting-edge AI development.
- Apply through OpenTrain
- Showcase specialized evaluation experience in your profile
- Grow a durable portfolio in AI training and data labeling
