Senior Agent Engineer – Production LLM Systems

About the role

We're looking for engineers who build and operate LLM agents in production — not in a notebook, not in a demo, but in systems that real users depend on and that you're personally on the hook for when they break.

This is a role for people who have real, hands-on visibility into how agents actually behave and get adopted inside an organization: what breaks, what gets ignored, what silently degrades, and what actually earns trust over time.

What you'll do

• Design, ship, and operate LLM-powered agents that real users rely on in production environments.
• Build evaluation systems and feedback loops to determine whether an agent got measurably better or worse after a change.
• Architect and maintain agents that outlive a single request — scheduled jobs, long-running workflows, and cloud sandbox execution environments.
• Instrument and monitor agent behavior in the wild, diagnosing failures, regressions, and edge cases under real usage conditions.
• Investigate and improve the layers most teams overlook: internal tooling, orchestration, memory/state management, and the operational glue that determines whether an agent actually gets adopted or quietly abandoned.
• Partner closely with stakeholders to understand adoption patterns — who uses the agent, who doesn't, and why.

Requirements

• Demonstrated experience shipping and operating LLM agents in production, with direct ownership of incidents and reliability.
• Strong understanding of agent evaluation methodology — you know how to measure whether an agent is actually improving.
• Experience building systems that run beyond single-request lifecycles: schedulers, background jobs, long-running or sandboxed execution.
• Familiarity with the full lifecycle of internal AI tooling, from build to real-world adoption (or lack thereof).
• Strong software engineering fundamentals and comfort operating with ambiguity in a fast-moving space.
• Excellent written communication for asynchronous, remote collaboration.

Compensation

This engagement pays $100–$500 per hour depending on experience and scope, fully remote, open globally.