Build production-grade agent systems that translate model capability into deterministic execution for real customer workflows.

Core agent harness: execution loop, tool-use strategies, context construction, and model-facing experimentation
Real workflow behaviors: define and refine agent behaviors across customer journeys and long-horizon tasks
Deterministic execution boundary: design the interface between runtime model decisions and deterministic, compiled, type-checked execution
Evals for release gating: run evaluations against replicas of real customer environments, reliable enough to gate releases
Production failure analysis: diagnose failures and drive systematic robustness improvements, attributed by layer
Computer-use in API-less environments: extend computer-use agents to environments without APIs, with production guarantees:
- per-task microVM
- model-blind credentials
- full action recording
- killable sessions
Feedback and data systems: implement loops and data systems to improve real-task data feeding into eval and training

4+ years of engineering experience
Experience with agent frameworks or tool-using LLM systems
Strong in Python or TypeScript, comfortable with modern AI tooling
Experience with model evaluation, fine-tuning, or prompt design
Ability to own systems end-to-end and debug across the stack
Systems thinking focused on user outcomes, not only model metrics
Comfort debugging messy, real-world failures and converting them into improvements
Motivation to work on the layer that turns model capability into systems that work for users
Can work in person in San Francisco, Monday to Friday, with startup intensity

Built computer-use or browser-automation agents
Experience with virtualization and sandboxed execution environments, including scaling
AI research experience with publications at top conferences
Shipped systems where correctness must survive partial failure (idempotency, resumability, compensating actions)
Integrated enterprise SaaS APIs (identity providers, directory services, ticketing, ERPs)
Experience with large, messy datasets or production logs
Early engineering experience involving direct customer communication

Senior/Staff Applied AI Engineer, Agent Harness

Job Description