Research
Research Engineer, Evaluation
- Location
- Remote (Africa, Europe or Americas time zones)
- Type
- Full-time
Define what "good" means for our AI products, and build the datasets and graders that prove it.
About the role
Every model, prompt and retrieval change at Radicand Labs has to pass an evaluation first. You’ll own that system: the datasets, the graders and the reports the whole team uses to decide what ships.
What you’ll do
- Build evaluation datasets from real, consented product usage
- Design graders: code checks, model-as-judge and human review
- Run experiments across models and providers and write up what you find
- Publish what we can on our blog
What we’re looking for
- Experience designing experiments and evaluating ML or LLM systems
- Strong Python or TypeScript
- Clear writing. Your results should change decisions
Nice to have
- Experience with code-understanding benchmarks
- Published research or open-source evaluation tooling