# We hire people who finish things.

How we hire, what we look for, and what production AI engineering here is actually like. No open role is listed on this page unless the role genuinely exists.

## Open roles

**No open roles right now.** We open roles when there is work that genuinely needs another pair of hands, and we say so here when that happens. If the description below sounds like you, write anyway — good people are worth a conversation ahead of a vacancy.

## Four things, consistently.

- **You ship to production, not to a demo** — The interesting part of this work starts after something works once. Retries, idempotency, evaluation, tracing, the failure paths — if that reads as chores rather than as the job, this will not be a good fit.
- **You can say what you do not know** — Most of what we build touches somebody’s real data and real money. Confident wrong answers are expensive, so calibrated uncertainty is a requirement rather than a personality trait we tolerate.
- **You read the system before changing it** — Almost every engagement is an integration with something older and stranger than you expected. The skill is understanding why it is the way it is before proposing that it should not be.
- **You write clearly** — Architecture decisions, handover documentation and client updates are all writing. A team this size cannot carry someone whose reasoning only exists verbally.

## Four steps, and you always hear back.

- **A written exchange** — You send something real — a repository, a piece of writing, a description of a system you built and what you would change about it. We reply with specific questions about it rather than a form.
- **A technical conversation** — Ninety minutes on a problem close to what we actually build: an integration with awkward constraints, or an agent design with a failure mode to close. No algorithm puzzles, no whiteboard performance.
- **A paid piece of work** — A scoped task of a day or two, paid at your rate. It is the only reliable signal either of us gets, and asking for free work from candidates is not something we are willing to do.
- **A decision, with reasons** — Either way you hear back with the actual reasoning. Silence after a candidate has spent two days on something is indefensible.

## How the days are shaped.

- Remote, with deliberate overlap rather than fixed hours.
- Small teams on one engagement at a time — no context-switching across three clients in a week.
- You own something end to end, including the handover conversation with the client’s engineers.
- Deep work is the default and meetings are the exception that has to justify itself.

Write to info@bralakai.com.

---

*Bralak AI — Building Intelligent Solutions · Automating the Future.* Bralak AI Pvt. Ltd. — Noida, UP, India.

- Canonical page: https://www.bralakai.com/careers
- Agent index: https://www.bralakai.com/llms.txt · full text: https://www.bralakai.com/llms-full.txt
- Contact: info@bralakai.com · https://www.bralakai.com/contact