Jobs at Deffai

This is a rare opportunity for a hands-on builder to work across the full product stack, build 0 → 1 and grow with a young team. You will work across the full product stack, build 0 → 1 and grow with a young team. From data pipeline, RAG, multi-agent orchestration, evals to human-annotation, you will get to experiment, explore solutions for longitudinal memory systems, LLM as a judge in eval tasks, and agent performance over diverse data sources. You will work on: 1. Collaborating with former FDA experts to accelerate their regulatory review workflows. 2. Improving RAG, citations, agent tool use, and memory across complex documents. 3. Fix a retrieval, citation, or tool-use issue in an existing workflow. 4. Add an evaluation case, measure the improvement, and ship the fix. 5. Build backend and frontend features using AI coding tools. 6. Debug agent workflows and improve quality, latency, and cost. 7. Turn expert feedback into product improvements and evaluations.