Model Behavior Engineer, Reasoning and Decision Making

The role

You will understand decision-making objectives and build evaluations to measure reasoning. You will design agent environments and software to improve how AI systems represent problems, use tools, and revise conclusions.

What you will do

• Identify user objectives, constraints, and available evidence
• Build evaluations for problems where success is underspecified
• Turn observed failures into testable hypotheses
• Experiment with problem decomposition, context, and tools
• Design software abstractions, data structures, and component boundaries
• Write maintainable code and participate in design and code reviews
• Integrate systems into the product

Required qualifications

• Rigorous thinking about ambiguous problems
• Curiosity across domains and clear communication
• Good experimental instincts
• Ability to independently write and debug code
• Ability to work with data and model APIs
• Proficiency in Python or a comparable language
• Practical familiarity with language models
• Strong engineering and architectural judgment
• Effective use of coding assistants

Nice to have

• Leveraged domain expertise across multiple scientific domains or business specialities

Benefits

• Flexible work environment
• Fully covered health, dental, vision, and life insurance
• Generous family coverage (FSA/HSA)
• Flexible vacation policy with recommendation of 20 days / year
• Company holidays
• $200 monthly wellbeing stipend
• 401K with a 6% employer match
• New Mac + $1,000 budget for workstation/home office in first year
• $500 annual budget for workstation/home office thereafter
• $1,000 quarterly AI Experimentation & Learning budget
• Team administrative assistant

About the company

Elicit helps research and business teams investigate hard questions and make high-stakes decisions grounded in evidence. They are building AI systems whose work people can inspect, question, and improve.