Backend Engineer

We run agent fleets in production to ship our own software, and we install the same capability in client teams. The backend that makes those fleets trustworthy is the part most of the industry skips, and it's the part you'd own. When an agent does the wrong thing at 2am, the systems you built are what catch it before a client does.

• 01Own the orchestration that decides who does what, in what order, and what happens when two agents disagree.
• 02Operate the context layer (our trovex) that serves the canonical answer about a codebase cheaply, so agents are fast and correct instead of confidently wrong.
• 03Make the fleet observable (our yoru) so we run it on evidence, not faith.
• 04Own services end to end on Postgres/Supabase and answer for them in production, pager included.

TypeScript/Node and Python, Postgres on Supabase, Next.js, and our own open suite (WRAI.TH, trovex, yoru). You'd operate the same stack we sell.

• 01

  You've run reliable systems in prod and carried the pager for them.

• 02

  You treat an LLM like production software, not a party trick.

• 03

  You'd rather instrument a thing than win the argument about it.

How we work: AI in production, not in a deck. We augment our people, never replace them — your craft stays yours, the output multiplies.