Senior Platform Engineer, Workflow Orchestration

We build a self-hosted tool for running workflows, the kind of thing teams reach for instead of standing up Airflow. We're hiring a senior platform engineer to own the scheduler at its core. It orders a graph of jobs, runs the ready ones with a bounded pool of workers, retries failures with backoff, and skips anything downstream of a failure. Concurrency is the heart of the job.

What you'll do:

• Run tasks concurrently through a worker pool without races, even though several workers touch the same results map and dependency counts
• Compute a valid execution order and catch cycles and broken dependencies before anything runs
• Propagate failures so dependents are skipped, and retry with backoff on a clock you can control in tests
• Handle cancellation cleanly, without leaking goroutines

What we're looking for:

• Strong Go concurrency: worker pools, mutexes, and code that stays correct under the race detector
• Solid graph algorithms, especially topological sort and cycle detection
• A bias toward correctness under load

Nice to have:

Worked on a scheduler or workflow tool, and comfortable with context cancellation and building for testability