What you'll do
- Get into a client's team — usually over a call, occasionally on site — map how the work actually happens today, and find the parts an AI system should take over, and the parts it shouldn't.
- Name the number each engagement is meant to move before we write any code, then track whether it moved.
- Turn that into something engineers can build from: flows, edge cases, data the system needs, and what "correct" looks like when it's an AI answering.
- Be the person who says a feature isn't worth the money. Most AI scope is bigger than the problem it's solving.
- Run the post-launch loop — what the system gets wrong in the client's real traffic, and what that means for the next iteration.
This role is fully remote, anywhere in India. Client discovery happens over calls, and the output is written — if you only do your best thinking in a room with a whiteboard, this will be harder than it looks.
What we're looking for
- 2-4 years as a BA, consultant, product analyst, or ops person who kept ending up owning the process redesign.
- You can interview someone about their job and come away with the workflow, not just the complaints.
- Comfortable with data — SQL or spreadsheets deep enough to check a claim yourself instead of asking someone to pull it.
- Clear, short writing. Our specs are read by four engineers, not filed in a portal.
- Enough understanding of what LLMs can and can't do to set client expectations honestly on the first call.
Nice to have
- CRM or sales-process experience — much of our work lands there, including Asobr.
- Having worked with clients outside India; ours are in India, the UAE and the US.
- Basic SQL against a Postgres database, or the willingness to learn it in your first month.