Senior Data Scientist, AI Retrieval Systems
Core
Build the retrieval and knowledge layer for AI systems supporting rare disease research at the NIH, including semantic search, ranking, and ontology management.
Role type
Senior individual contributor machine-learning engineer (retrieval systems)
Builds
Retrieval-augmented services, knowledge layers, and interfaces for biomedical research
Domain
Biomedical informatics / Rare disease research
Deliverable
production ML models | product features
Required skills
LLM application engineering, retrieval system design, vector search, PostgreSQL, Kubernetes, Next.js/React/TypeScript, evaluation of open-ended systems, structured output
Preferred skills
Biomedical ontologies (MONDO, HPO, UMLS), entity linking, on-premises HPC deployment, open-weight model serving, rare disease domain knowledge
Technologies
Python, FastAPI, Pydantic, pytest, PostgreSQL, pgvector, Kubernetes, Helm, Next.js, React, TypeScript, Ollama, vLLM, LiteLLM
Responsibilities
Ingest and reconcile disease/phenotype ontologies into PostgreSQL; build retrieval services grounding everyday language in clinical concepts; tune keyword and vector search over biomedical corpora; build ranking and relevance layers; develop user interfaces in Next.js/React/TypeScript; deploy services to NIH on-premises Kubernetes environments; build evaluation regression suites; log system behavior and decisions; collaborate with researchers to define data models and interface design; contribute to scientific publications.
Seniority
Senior, hands-on IC
