5 months ago
Los Altos, CA, USAMid Level
Responsibilities
- Design, build, and operate the agent runtime, harness, tool-calling infrastructure, sandboxed code execution, memory, reliability, and observability layers.
- Build backend services and APIs using Python, FastAPI, and Postgres.
- Operate vector databases, semantic search, and RAG systems while improving search quality and latency.
- Develop ingestion, cleaning, and indexing pipelines for academic, clinical, regulatory, safety, and patent data.
- Build systems that ground AI-generated outputs in source data with traceable, verifiable citations.
- Work with founders, domain experts, and early customers while owning outcomes in an ambiguous startup environment.
Requirements
- At least 2 years of experience shipping and operating production backend systems end to end.
- Strong understanding of API design, data modeling, databases, and system reliability.
- Hands-on production experience with LLMs, RAG, and vector databases.
- Practical experience building or deeply tracing agent runtimes, harnesses, tool calling, sandboxing, and memory systems.
- Ability to set priorities, ship quickly with incomplete specifications, and balance speed with engineering judgment.
- Active use of AI coding tools such as Claude Code and Cursor.
- Preferred: experience building an in-house agentic harness or productivity tooling, operating LangGraph or other agent frameworks and RAG systems in production, building enterprise SaaS or pharma, biotech, or healthcare products, working with life sciences data, deploying production systems on AWS, Terraform, and Kubernetes, and working at a seed or early-stage startup.
Benefits
- 401(k) retirement plan.
- Medical, dental, and vision health insurance.
- Lunch and dinner meal allowance.
- Transportation support for early starts and late nights.
- In-office snack bar plus additional commuting and work travel support.
- The role is based in the office in Los Altos, California; the interview process includes screening, a take-home assignment, a technical interview, and a cultural interview.
Tech Stack
Categories
