Our client is building the data layer for a world where AI agents, not humans, run production workloads. Today's data platforms assume a careful person at the keyboard. This one lets agents run, inspect, verify, and undo real jobs against production data, safely. They already have paying customers using it live.

You'd join as one of the first engineers building this piece: real ownership over how it gets built, not just a seat on someone else's roadmap. Specifically, you'd own the Runner, the layer that executes compute at scale in the cloud. Python is the main language, backed by a growing amount of Rust.

This is a hands-on tech lead role. You make the calls on your area and stay in the code full time; there's no path to management here, by design.

What you'd work on

• Systems that let people and agents safely run, check, and roll back data jobs at production scale
• APIs and abstractions simple enough for a human to read and solid enough for an agent to call
• Performant execution of Rust, Python, and SQL workloads over object storage
• Runtime work: scheduling, concurrency, caching, reliability

What you need

• 7+ years building and running distributed systems
• Fluent in Python plus a low-level language; real Rust experience is a strong plus
• A track record of owning ambiguous problems from first requirements through production
• Comfortable writing design docs and leading technical discussions

Good to have: object storage and data lake internals, query engines like Arrow, DataFusion, or DuckDB, distributed compute (Kubernetes, Spark, or similar), or work building infrastructure for AI agents themselves.

Base pay matches the market rate for the right candidate, plus 0.40-0.60% equity with a standard 4-year vest.