Model Behavior Engineer, Reasoning and Decision Making

Engineer AI systems that help teams make high-stakes decisions by improving model reasoning, evaluation, and legibility.
- Understand real decision-making contexts with users and domain experts; identify objectives, constraints, and evidence gaps.
- Design evaluations for complex reasoning tasks; combine expert judgment with quantitative analysis to measure meaningful improvements.
- Improve agent behavior through experimentation with problem decomposition, tools, context, and information-seeking strategies.
- Build maintainable software for reasoning workflows with clear abstractions, reliable failure modes, and integrated monitoring.
- Requires strong Python skills, practical language model experience, and rigorous thinking about ambiguous problems.

Elicit is a public benefit company developing an AI research assistant to help researchers make better decisions. Their product aims to be a scalable ML system prioritising systematicity and transparency, with supervision of process, not outcomes.