A generalist engineering track covering backend services, data pipelines and web surfaces. You'll be handed problems, not tickets, and you'll be expected to come back with a solution.
Every company sees one part of your life. Alone, each is an incomplete picture. Together, they can answer questions no single company can. A student applying for their first credit card gets rejected without a credit history. But years of rent, bills and shopping tell a different story.
Companies can't just hand each other their data.
It's private, regulated and their most valuable asset.
LattIQ solves exactly this.
A platform to let companies learn from each other's data without it ever leaving its boundaries. We're building it from scratch, and some of India's biggest financial companies are already using it.
We're a small team, founded by people who've built at Microsoft, Flipkart, Grab, CRED, and Elevation Capital. You'll work directly with us. No layers, no bureaucracy, real ownership from day one.
You'll ship features that paying clients use.
Founders accessible for a discussion at all times
Too small for anyone to get boxed in. You'll have shipped across the stack by the end.
Strong interns are first in line for full-time offers as the team grows.
Ship a small end-to-end change.
Own a feature.
Reviewing other people's code.
The core services the platform runs on: endpoints, authorization, storage, throughput. You design from what the system must never do. Raw data behind an insight can't reach the other client, or us.
Jobs that ingest and validate data, then turn it into scores and audiences. They run in a client's environment, over data you'll never see. You debug from metrics, not print statements.
You deploy what you write, to two places. Our platform runs on AWS. What ships to a client runs on infrastructure we don't own, so it has to be portable.
Screens on our design system, used by analysts and data teams at large financial institutions.
From rough ask through design, tests, review and release, then stay responsible once it's live. Schema changes, client data and anything expensive to undo get a founding engineer's sign-off. The rest is yours to call.
Weekly iterations, and priorities move. Scope your own work, flag blockers the day you hit them. When something won't fit the week, cut the scope, not the bar.
Use LLMs daily; our repos carry context files so the tools have something real to work with. The review bar doesn't change: if you can't explain it, it doesn't ship.
Unfamiliar technology most weeks. Close the gap yourself and bring sharp questions. Help on judgment and design calls, yes. A curriculum, no.
Tests and documentation are part of "done". There is no separate intern bar.
Instrument it, set the alert that tells you when it breaks, and trace failures to a root cause. Production behaviour is your problem.
Done by hand three times, costly to get wrong, or leaving another team waiting on an engineer: it belongs in a script or a self-serve surface.
You'll be in design reviews and customer conversations. Argue a position, and change your mind when someone has the better argument.
You can pick the right data structure and defend the choice, and you spot the code whose cost grows faster than the data does. We'll ask design questions as well as puzzle ones: what you'd trade off, what you'd measure, what you'd do differently at ten times the scale. Your problem count is not the signal.
We'd rather you be strong in one than shallow in five, and we don't mind which one. It has to be a language you've used to build something that runs on a server: concurrency, memory, failure paths, state that survives a restart, what breaks when traffic multiplies. Frontend-only experience won't clear this bar by itself. Interview in whichever language you know best; you're not expected to arrive knowing ours.
A good share of what you build ends up on a screen, and you'll carry a feature all the way to the user. Not design talent or framework mastery: HTML, CSS and JavaScript, how a component renders and re-renders, how a browser talks to a backend. We build on React and TypeScript, though you don't need to know them already. Depth on the server, competence on the client.
We won’t ask you to code without them. We’ll ask you to explain what you shipped: "the model wrote it" is not an answer to "why is this line here". Reading code you didn’t write is the single best predictor of a good intern here, because most of the job is understanding an existing system before adding to it.
We write our decisions down. PR descriptions, design notes and ADRs get read long after the conversation that produced them.
Starting immediately. This is not a part-time internship. Preferred to work in-person directly with the founders
None of these are filters
Version control, code review, automated tests, CI pipelines. Almost nobody picks this up in a degree programme, so its absence costs you nothing here.
You can write a non-trivial join and explain why an index matters.
A side project, a college project you finished, an open-source contribution, a hackathon build, or a write-up of a bug you chased down.
A cloud account you've used, a container you've built, a pipeline you've set up.
How models get built and evaluated, or how cryptography and privacy engineering make data useful without exposing it.