- Own one AI, ML, or computer-vision product end to end: problem, users, roadmap, release.
- Set the KPI a release has to move, and hold the gate when it does not transparent it.
- Decide what ships, what waits, and what dies.

Working with engineering

- Write specs that survive the codebase, the kind engineers trust.
- Reason about architecture, cost, and latency well enough that no one can bluff you.
- Prototype your own ideas with AI tools instead of delegating every question. The hard part, the part most product managers dodge

- Design for the failure mode, not just the demo: the edge cases, the false positives, the long tail, the last twenty percent.
- Ship against evals and real numbers, not a spec that is either met or not.
- Build the guardrails and the fallback that turn a demo into something an enterprise will actually run. Outcomes, not output

- Tie the work to a number the business actually feels.
- Turn what works once into something repeatable.
- Say plainly what is on track and what is slipping, every week. This role is not for you if





- You need the product handed to you before you can care about it.
- You think generative AI is the whole of AI.
- You count features shipped instead of the number you moved.

- You go quiet at eighty percent instead of owning the last twenty.
- You need a title to feel credible, or you think a title is enough.
- You want predictable, low-stakes, roadmap-and-standups work.

How to apply

- Apply through the steps that follow. The questions are built to surface how you think and what you have built, not how well you polish a resume. AI-written answers are obvious, and they are rejected.

- Everything you write, we will dig into if we talk.

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change.

Applicants are advised to confirm the latest information directly on the company website before applying.

📌 Product Manager, AI / ML / Robotics (Mumbai)
🏢 Allerin Tech
📍 Mumbai