Engineering Manager

Required

• How you lead. You manage people, not tickets. You know what each engineer on your team is trying to grow into, you give direct feedback early, and you’ve had the hard conversations: the engineer who’s checked out, the promotion that isn’t ready yet, the call that upsets someone. Strong engineers want to work for you because you make them better and get out of their way.
• How you deliver. You turn half-shaped problems into a sensible sequence of work, spot the project that’s quietly slipping before it slips, and know when a plan needs more structure and when it needs less. You have the system design chops to help an engineer through a hard design, and to tell when an estimate or a blocker doesn’t hold up.
• This isn’t your first EM role; you’ve hired, run performance cycles and grown people into bigger scope.
• Technically strong: at least a solid senior engineer by our bar. You don’t need to be the best engineer on the team, but you’ve built and run production systems yourself, and the engineers you lead respect you because you’ve done the work.
• Live in an agentic coding harness (Claude Code, Codex, Cursor — we don’t care which), with real opinions about how a team, not just an individual, should work with agents to ship fast without letting quality slip.
• Resourceful under ambiguity. You add just enough process to keep a team unblocked, and no more.
• Close to the product. You care whether what shipped actually worked for customers, not just whether it shipped.
• Write clearly. We’re remote and async-heavy; people who write clearly think (and prompt) clearly.
• We’ve rebuilt how we work around AI agents, so our engineers are unblocked to work at AI-native speeds. Part of your job is keeping it that way as the team grows.
• Main goes to production and approvals aren’t always required. Fast, automated CI is the gate, and we trust people to make the right judgments.
• We invest in making the codebase easy to work in: written runbooks and conventions that help agents and new people find their way around quickly.
• Running a few agents in parallel is easy, and expected. Sandboxes and preview environments are trivial to spawn.