We're building autonomous agents that handle end-to-end customer operations. You'll design tool-use interfaces, build planning loops, and own the eval harness that keeps our agents reliable.

What you'll do: - Build production agents with robust tool use and memory - Design and grow our evaluation suite - Partner with product to ship agent features customers trust

What we're looking for: - 4+ years software engineering, strong Python - Hands-on experience wiring LLMs to real tools and data - Comfort with non-determinism and eval-driven development