Founding Harness Engineer – AI Agents

Job Description:

• Build the agent runtime, including context, tool choice, approval waits, and decisions about when to act, ask, or stay quiet
• Write and continuously tune system prompts
• Own the eval and regression harness and test prompt changes like code
• Make long-running work resilient to interruptions, retries, and flaky APIs
• Implement guardrails, approval gates for risky actions, and strict separation of customer data
• Work directly with the founder and early users and turn feedback into shipped behavior

Requirements:

• Shipped tool-using LLM agents to real users in production; demos and prompt wrappers do not count
• Good instincts for agent initiative, supported by test results
• Hands-on experience building eval and regression suites for LLM behavior
• Understanding of safe tool execution, prompt injection, and multi-tenant data risk
• Solid backend skills in Python or TypeScript, including async jobs, retries, and state
• Ability to take a fuzzy problem from prototype to production and explain the tradeoffs
• Excellent written and verbal communication
• Nice to have: MCP or integrations with Slack, Gmail, Notion, or Linear
• Nice to have: eval and tracing tools such as Braintrust, Langfuse, or LangSmith
• Nice to have: experience as an early engineer at a startup

Benefits:

• Competitive compensation + equity
• A founding seat with ownership of the agent harness and influence over product behavior
• A direct line to the founder and early users, with no layers in between
• Remote anywhere in the US, ideally keeping hours that overlap with Pacific time
• The tools and models needed to do the job properly
• A small, experienced team that trusts you to run your own work
• Path to full-time once it's clearly working for both sides