Founding ML Engineer
First seen by Alion on Sep 7, 2026.
The role :
You will own the full ML lifecycle : the data engine, training, evaluation, deployment to production, monitoring in the wild, and the loop that ties them together. Real factories are the hardest possible environment for computer vision -
bad lighting, blocked views, camera angles you didn't choose, hundreds of subtly different processes - and your
models have to be right, in real time, with customers watching.
Who we're looking for :
Applied-first, with genuine research depth. You've trained models yourself - not orchestrated APIs - and you've also carried them all the way into production and lived with the consequences. You read papers, implement the ones that matter, and know when the frontier idea beats the boring baseline and when it doesn't.
Strong signals :
- You've owned an ML system end to end in production : data, training, evals, deployment, monitoring - and iterated it against real-world failure, not benchmark leaderboards.
- Deep computer vision fundamentals : you can reason from first principles about why a model fails on real video, and design the data or architecture change that fixes it.
- You've built or run a serious data engine - collecting data, deciding what to label, hunting down the hard examples, building test sets that actually predict real-world performance.
- You did your best work at an early seat in a high-performing startup : small team, no safety net, outcomes on you.
- Evidence of exceptional work anywhere : published research with real impact, a system people said couldn't be built, a result that made your previous team's product possible.
- We care very little about keyword coverage. We care whether you've done something extraordinary and can explain precisely why it worked.
Who this is not for :
- If your ML experience is prompt engineering, RAG pipelines, or fine-tuning through an API - this isn't your role.
- If you've done research but never shipped a model that real users depended on - this isn't your role.
- If you're optimizing for work-life balance : we work 6 days a week, in person, in Gurgaon. The pace is intense and the bar is high. This is a build-a-company job, not a lifestyle job.
What you get :
- Industry-leading pay + meaningful founding equity (up to 1%)
- Full ownership of ML at a company where ML is the product, working with the founders directly
- Fast, real feedback : every model you improve shows up on factory floors within days, not quarters
Skills
Computer Vision, Machine Learning, MLOps, Prompt Engineering