Product Engineer, AI Agent

The role

Own how an agent performs its work: the information it receives, the tools it can use, and the points where it needs a person. You’ll turn a defined workflow into software we can test, observe, and improve after release.

What you’ll do

• Build agent workflows that use client systems through well-defined tools and APIs.
• Design state handling, retries, and recovery paths so interrupted work can be understood and resumed safely.
• Add approval steps and permissions appropriate to the actions an agent can take.
• Create evaluations from real workflow examples, then use traces to investigate failures and regressions.
• Work with product engineers on the review experience and extract reusable components from repeated work.

What you bring

• Experience shipping and maintaining software used by other people.
• Hands-on work with an LLM application, including tool calls, structured outputs, or retrieval.
• Strong backend skills in Python or TypeScript and confidence integrating external services.
• A habit of testing failure cases and measuring quality, latency, and cost before changing a system.

Experience with queues, background jobs, evaluation datasets, and production observability would be useful.

About Olypsis

Olypsis identifies where AI can grow a business and builds custom agents to execute its playbooks. We work with clients to understand the process, build the software, and keep improving it once it is running.