Staff Engineer, AI
LiteLLMSan FranciscoOn-Site
Needs a free Superjobs account. You'll finish on LiteLLM's Ashby page.
Skill score
Summary
The Staff Engineer leads the unified API team, owning rollouts of new models, endpoints and providers through LiteLLM's gateway, triaging open source translation issues, and reviewing code from the AI team. The role suits a hands-on engineer with applied ML and LLM API depth who can set technical direction and work closely with customers.
- Function
- Development
- Estimated salary
- Est. $156k–234k/yr
- Estimated for a staff-level role, ~3 yrs experience, United States. Not published by the employer.
- Work mode and location
- On-Site, San Francisco
- Experience
- 3+ yrs
- Employment type
- Full-time
Job description
Overview
We're looking for a Staff Engineer to lead our unified API team and make sure every new model and provider works with LiteLLM on day one.
About the role
This is a hands-on technical lead role. You'll own how LiteLLM unifies LLM APIs across the OpenAI and Anthropic specs, lead the engineers who build it, and set the bar for what good looks like.
Every week a provider ships something new: a new API, a new modality, a new feature. The promise of LiteLLM is that it works with all of them on day one. Your job is to keep that promise: new endpoints ship on day one, open source translation issues go down, and developers can count on LiteLLM to keep up.
About the team
You’ll lead our open source pod, where most issues are LLM translation issues. You’ll report to the CEO, and later to our Head of Engineering. This role is in person in San Francisco. We help with relocation and support H-1B transfers.
Roles and responsibilities
- Lead the unified API team and own new model, endpoint, and provider rollouts end to end.
- Design provider-agnostic features at the gateway level, such as session management for non-OpenAI models.
- Triage open source issues and outside PRs (new providers, translation fixes) and get them merged, reviewed, or replaced with an internal fix.
- Review the AI team's PRs and hold a high bar for code quality, including code written with AI.
- Find the power users in your area and use what they need to decide what to build next.
- Take on new unification bets with a customer who will use them and give feedback.
- Own the developer experience for everyone who uses LiteLLM's Python SDK and proxy.
Skills required
Top skills
- Python
- LLM APIs
- Unified API design
- OpenAI API spec
All skills (8)
- Anthropic API spec
- FastAPI
- Redis
- Postgres
- Rust
- LLM evals
- Open source contribution
- Developer experience
Qualifications
- You've been a machine learning engineer and now build applied AI products.
- 3+ years of experience. The scope you've owned matters more than a Staff title.
- Deep expertise in LLM APIs. You can speak to the differences between /chat/completions and /responses, and to each provider's nuances.
- Experience building or maintaining a unified API that sits in front of many providers.
- Taste in developer experience. You can tell when an API or playground is bad and explain why.
- Customer empathy. You ask why a user wants something and how big the use case is before you build it.
- Ready to lead a team from day one and set the technical bar yourself.
- Strong software engineering fundamentals in Python.
Nice to have
- Experience at a company built on a unified API, such as OpenRouter, Vercel AI SDK, Merge, Segment, or Fivetran.
- You've built your own unified API or written about what a good one should do.
- You've built evals for LLMs.
- You've used LiteLLM in production.
- Open source contributions, especially to LLM tooling.
Other details
The tech stack includes Python, FastAPI, Redis, Postgres. We're moving the gateway to Rust.
Hiring process
If you want to lead the team that keeps LiteLLM ahead of every model launch, we'd love to meet you.
Company info
About LiteLLM
LiteLLM is the world's most widely adopted AI Gateway. We provide a unified API for 100+ LLM providers along with routing, authentication, budgets, observability, and governance for production AI systems. Thousands of companies rely on LiteLLM to power AI in production.