About the role
You will design and run the automated and semi-automated test batteries that power A.E.G.I.S.™ evaluations across all seven safety pillars. Your work directly determines the ratings the market relies on.
What you'll do
- Build evaluation harnesses that probe AI systems for hallucination, guardrail failure, and privacy leakage
- Develop reproducible test suites and scoring pipelines
- Translate messy real-world failure modes into rigorous, repeatable tests
- Partner with the red team to turn discovered exploits into standing checks
What we're looking for
- Strong engineering background with experience testing ML or LLM systems
- Comfort designing tests for systems that behave non-deterministically
- A rigorous, skeptical mindset and clear technical writing
Why it matters
The systems you evaluate are headed for children's bedrooms, hospital workflows, and financial decisions. Your tests are the thing standing between a safety claim and the truth.
Placeholder job description — replace with the final posting.
