AI Engineer
- Compensation
- $175K–$190K / yr
- Location
- San Francisco
- Work setup
- Hybrid
- Seniority
- Not specified
- Employment
- Not specified
- Experience
- 3+ years
- Category
- Ai engineering
- Posted
Skills
Must have
- 3+ years of experience in AI/automation engineering, workflow automation, or a technical operations role, ideally at a high-growth, AI-native infrastructure or B2B company
- You've shipped production agents on Vercel. Build custom internal apps and agents on Vercel, with durable workflows and sandboxed execution, so what you ship runs reliably in production instead of as
- Genuine fluency with AI coding assistants and agent tooling (Claude Code, Cursor, Codex, or similar)
- Direct experience with a CRM or system-of-record platform (Salesforce or similar)
- A track record of owning end-to-end AI and automation workflows
- Proficiency with integration and automation platforms (n8n, Zapier, Make, Workato, or similar) and a fundamental understanding of APIs and webhooks
- Experience in high-growth technology companies, ideally in infrastructure, cloud, or operations-heavy environments
Nice to have
- Familiarity with GPU infrastructure, capacity planning, or fleet management concepts
- Comfort building reporting and alerting on a data warehouse (BigQuery, Databricks) and BI layer (Sigma, Hex) — not as a data engineer, but as a consumer and builder on top of the data layer
- Experience building with agent platforms (Gumloop, n8n, Notion Agents, etc.) and/or agent frameworks (Vercel AI SDK, Claude Agent SDK, Mastra, etc.)
- Experience with high-stakes, operationally complex environments where mistakes have real business impact
- Challenging work and exposure to the operational core of a fast-scaling AI infrastructure company
Full description
ABOUT BASETEN
Baseten powers mission-critical inference for the world's most dynamic AI companies, like Cursor, Notion, OpenEvidence, Abridge, Clay, Gamma, and Writer. By uniting applied AI research, flexible infrastructure, and seamless developer tooling, we enable companies operating at the frontier of AI to bring cutting-edge models into production. We're growing quickly and recently raised our $1.5B Series F, led by Altimeter Capital, Conviction Partners, and Spark Capital. Join us and help build the platform engineers turn to ship AI products.
THE ROLE
Baseten's Compute org is in hyper growth. As it scales, the systems and workflows that keep supply and demand balanced across our GPU fleet need to get more sophisticated, and this role exists to make sure they do.
Compute sits at the center of how Baseten allocates, forecasts, and manages the capacity that powers every customer inference request. The team that supports this work, C3, runs on a mix of internal tooling, manual processes, and systems that haven't fully kept pace with the scale of the problem. This role exists to close that gap.
You'll design, build, and ship AI-powered workflows that give the Compute and C3 teams real leverage, automating the manual, repetitive, and error-prone parts of the capacity lifecycle so the team can focus on judgment calls that actually need a human. We want someone who can walk in, audit what exists today, identify what's missing or broken, and start shipping fast.
You know when to reach for an existing internal tool and when to build something custom in Claude Code. You think two to three steps ahead about how the thing you build today fits into the broader capacity systems architecture tomorrow. And you bring a point of view on our stack, on what we should be building, and on where AI can do something existing tooling simply can't.
RESPONSIBILITIES
- Ship AI-powered workflows for Compute and C3: build the agents and automations that give capacity analysts, ops leads, and engineers real leverage, off-loading manual and repetitive work like data pipeline cleanup, migration troubleshooting, and capacity investigation.
- Get insights in front of the team: turn fleet utilization, allocation, and demand-forecasting data into the dashboards, alerts, and recommendations that C3 and Compute leadership actually act on. A build isn't done until the team is using it.
- Audit the stack and generate your own backlog: Compute runs on a mix of tooling with real overlap and real gaps. Walk in with a point of view, identify what's missing or broken, and prioritize without waiting to be handed a roadmap.
- Ship team-productivity workflows fast: triage asks from Supply, Demand, and C3, find the low-hanging fruit, and build it. A good week looks like an ops lead asking for something Monday and having it live by Wednesday.