Build the models behind every delivery
We’re a small team in Mumbai training our own routing, forecasting, maintenance and vision models on real logistics data.
Mumbai, India
Why this is worth your time
You'd train real models
Not prompt-engineering over someone else's API. Graph neural networks for routing, quantile forecasting, telemetry models, warehouse vision.
The problem is physical
Your model output becomes a truck taking a different road. Feedback is concrete and measurable against a real baseline.
Small team, wide scope
Early enough that you'll own whole problems end to end rather than a slice of someone else's design.
Built in Mumbai
Indian logistics is one of the largest and least optimised networks in the world. That's the advantage, not the compromise.
Open roles
We don’t have listings published right now, but we’re always interested in engineers who’ve trained models on messy real-world data. Send a general application below and we’ll come back to you.
How we hire
Four steps, usually inside two weeks.
You apply
Send a CV or a link to something you've built. We read every application ourselves — no keyword screen.
Intro call
Thirty minutes on what you've worked on and what we're building. Honest on both sides about stage and scope.
Technical conversation
A real problem from our domain, discussed together. No whiteboard algorithm puzzles unrelated to the job.
Offer
We move quickly. If it's a fit you'll know within a week of the technical round.
Apply
A link to something you’ve built counts for more than a polished CV.