AI safety & policy
Part-time
Routes to impact
Skill-building & building career capital
Direct high impact on an important cause
Description
Build evaluation software for AI models and agents, testing behavior changes and producing evidence for research and product workflows.
- Develop evaluation harnesses for model, tool-use, and multi-step agent behavior
- Record inputs, tool calls, observations, and limitations in structured traces
- Build replay and regression tests to compare behavior before and after changes
- Document scenarios, run instructions, results, and evaluation limits
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.



