AI Engineer (Clinical AI Data + Retrieval) — San Francisco (On-site) | $170–250k + equity
I’m working with a well-funded SF healthcare startup building a consumer primary care product that combines patient records + AI + real clinicians to deliver fast, explainable, citation-backed medical answers.
They’re hiring an AI Engineer to own the data + retrieval foundation behind the clinical AI. This is not prompt tinkering - it’s systems work: pipelines, traceability, correctness, observability, and performance.
What you’ll do
- Build/improve a clinical knowledge base where citations resolve to original sources
- Improve retrieval quality (RAG / grounding / vector search) with strong data lineage
- Rebuild data ingestion + integrity to eliminate stale data & support corrections
- Implement observability for background jobs (tracing, failure analysis, auditing)
- Design data portability + deletion workflows with privacy built-in
Must-haves
- Strong Python + SQL
- Production Postgres (JSONB; vector is a plus)
- Proven refactors / migrations / simplifying messy data systems
- Comfort owning problems end-to-end (schema → service → prod debugging)
- Good judgment around correctness and tradeoffs
Nice-to-haves
- RAG pipelines, vector search, eval frameworks
- FastAPI / SQLAlchemy / DuckDB / Temporal / ClickHouse / Valkey
- Langfuse/Sentry-style observability + Docker/ECS deployments
- Experience with regulated or sensitive data (bonus, not required)
Package
- $170k–$250k + equity (level depends on ownership/scope)
- On-site in San Francisco (non-negotiable)
