Senior AI Infrastructure Engineer · San Francisco Bay Area
I design and ship production AI infrastructure — systems where model routing, agentic loops, safety controls, and cost budgets are first-class concerns, not afterthoughts.
Current focus areas:
- LLM routing — task-aware multi-model dispatch, cost optimization, fallback chains
- Agentic architecture — Observe-Decide-Act loops with durable audit trails and explicit safety gates
- RAG at scale — multi-hop retrieval, GraphRAG, hybrid dense/sparse pipelines
- Health-AI safety — scope enforcement, PII de-identification, escalation triggers
| Project | What it does |
|---|---|
| nexus-llm-router | Drop-in OpenAI-compatible API that routes requests across LLM providers with task-aware strategies, per-user cost budgets, and durable audit logs |
| scholar-rag-agent | Production-grade Agentic RAG for scientific literature — multi-hop reasoning, GraphRAG, multi-LLM routing |
| medagent-core | Safety-first clinical AI reasoning agent — explicit state machines, Bayesian hypothesis ranking, mandatory disclaimers |
| multi-bot-agentic | Deterministic multi-provider AI-agent orchestrator with rationale traces, safety controls, and durable event logs |
| agentic-career-search | Autonomous AI-agent orchestration engine for job discovery with decision traces, tool adapters, and production-grade run control |
Building infra that lets teams ship reliable AI — without surprises in production.