HomeCompanyCareers
CompanyBuild whatmatters.
A small team. Real ownership. AI that reaches credit desks, claims teams and the people they serve.
- 01 · DataPipeline
- 02 · ModelEvaluate
- 03 · ImpactShip
01How we operate
What it’s like to work here.
Small team, direct ownership, real customers from week one.
You’ll sit with the credit officers, adjusters and clerks who use what you build. The problem is never theoretical.
From data pipeline to eval harness to the inference running in a customer’s VPC — you ship the thing end to end.
We measure everything and we’d rather say ‘not yet’ than ship a confident guess into a regulated workflow.
On-site by default because the work is collaborative and fast. Some roles are remote within India.
02Open roles
Where we’re hiring.
Don’t see your exact role? If you do work that’s relevant to regulated AI, write to us anyway.
03How we hire
A short, honest process.
Four steps, real work, no trick questions. Usually wrapped up in two to three weeks.
- 01
Intro call
Intro
A conversation about your work and what you’re looking for — and an honest read on fit.
- 02
Technical deep-dive
Deep-dive
We dig into something you’ve actually built. No whiteboard puzzles.
- 03
Working session
Session
A scoped exercise close to the real work, done together with the team. It is paid, and we agree the scope and the time it should take with you before you start.
- 04
Team & offer
Offer
Meet the people you’ll work with, then a decision — fast, either way.
04Questions
Common questions.
Most engineering roles are on-site in Bengaluru because the work is collaborative and client-embedded, but some roles are open to remote within India. Each listing says where it’s based.
Yes, for roles where it makes sense. We’ll discuss specifics during the process.
Python-heavy ML and data tooling, modern inference infrastructure, and product surfaces in TypeScript/React. We train and fine-tune our own models.
If you’ve shipped something real and you’re drawn to high-stakes problems, yes. We weight evidence of building over years of experience.
Join us
Tell us what you’ve built.
Send the role you want, a CV or LinkedIn, and one thing you’ve built that you can talk through in detail — a repo, a paper, a system in production. The last one carries the most weight.