Questions about the Sprint.

A 4-week project cohort. You build one production GenAI system with a small group: a deadline every week, automated checks on your code, reviews from classmates, one live session a week, and a demo at the end. There are five Sprints, one system each.

Because each one ends with something real and finished. Five systems cover what a GenAI engineer builds at work: a RAG system, an AI agent, a fine-tuned model, evaluation and guardrails, and deployment with monitoring. Four weeks is long enough to build one properly and short enough to finish.

No. Each Sprint stands on its own, with its own price. Join the ones you need.

Sprint 01, the RAG system, runs first, and most GenAI work starts with retrieval. If you already ship RAG systems at work, wait for the Sprint that covers what you don't.

Once the waitlist is big enough for a pilot group. Everyone on the waitlist hears first, by email.

They get upgraded as the tools change, about once a year, so you build with what companies use now, not what they used when a course was recorded.

No. You need Python and some data science or software background. The free path covers the ideas you'll use.

Start with Python and the free path first. The Sprint moves fast, and you'll get far more out of it once you can read and write Python comfortably.

Plan for about 5 to 6 hours a week, including the live session.

Yes. For help with your own situation now (a career move, a RAG system, your resume or an interview), book a 1:1 call. It's separate from the Sprint, and you don't need it to join.

You push one deliverable to your own GitHub repo by the deadline. Your repo is run against evaluation scripts, so you see scores, not just a tick. Classmates review your pull requests and you review theirs. Then there's one live session with Sumit: questions, code reviews, and the week ahead.

During the Sprint you test your system on questions you can see. In week 4 it's also scored on hidden questions you never see. You can't tune your system to questions you've never seen, so the score shows your system actually works.

Three layers: the automated checks on every push, your classmates on every pull request, and Sumit in the weekly live session.

No. You get a starter repo and a brief, and you build the system yourself. That's the point: at the end it's your work, and you can explain every decision in an interview.

Yes, use any tool you like. The checks don't care who typed the code. They care whether your system is accurate, fast and cheap, and whether it holds up on the hidden tests. Using AI well is part of the job now.

AI can write the code. It can't make you finish, it can't give you proof a hiring manager trusts ("the AI said my code is good" isn't proof), and it answers the question you ask, not the one you should have asked. The Sprint gives you deadlines, a hidden-test score from an independent judge, and reviews from people who ship these systems.

One flat, low price per Sprint, announced to the waitlist first. The waitlist form asks what you'd pay, and those answers set the price.

You don't pay for information: every video, guide and path stays free. You pay for what free content can't give you: deadlines and a group so you finish, scores and reviews on your own code, and a finished project with proof. Like a gym: the workouts are free online, and people still pay for the trainer.

You finish with a public repo, an evaluation report, a hidden-test score and a demo. A hiring manager can check all four, which is more than a certificate gives them.

No. Nobody can honestly promise you a job. What you get is real work you can show and explain.

No. No upsells, no sales calls, no bundles.

The refund rule is published together with the price, before anyone pays.

Yes. Follow the free path. The Sprint adds deadlines, reviews and a group to build with.