Programs
Two clubs, and whatever else we can support.
ML Club and Build Club run continuously. Everything else — hackathons, study jams, sessions — runs when someone makes it possible.
Running now
The two clubs
These don't depend on sponsorship. They're the community itself.
ML Club
Math Club + Paper Reading Club
Two tracks running side by side. The Math track builds the foundations AI actually rests on — linear algebra, calculus, probability and statistics, optimisation — working through Khan Academy, 3Blue1Brown and Stanford CS229 together rather than alone. The Paper Reading track picks AI papers, reads them as a group, and each member explains what they understood on camera or in writing.
Math Club
Linear algebra, calculus, probability and statistics, optimisation. Roughly three to four months to cover all four, taken one at a time.
Paper Reading Club
Pick papers, read as a group, then present what you understood. Sessions are recorded and the recording goes out on your channels, not just ours.
What you get
- Medals, swag and t-shirts when you hit the milestones
- Your session recorded, and shared on your own socials
- A group that reads the hard parts with you instead of leaving you to it
- Something concrete to point at — a talk, a write-up, a recording
- For
- Anyone who wants the fundamentals to stop being a gap. No registration, no prerequisites — you can join at any point in a cycle.
- Join
- Join the group and come to the next session
Build Club
Build, share, get feedback
Nothing fancy — a place where people build things with AI or build AI, post what they made, ask for feedback, and share the resources that helped. It's what keeps momentum going between events.
What you get
- Honest feedback from people who've shipped something
- Collaborators for the thing you can't finish alone
- The resources that actually worked, not a link dump
- For
- Anyone building something, at any stage, however rough.
- Join
- Post what you're working on in the group
We are a volunteer community and nobody takes a salary. Programs are free whenever we have support. Where support isn't there, we may charge only what the session costs to run — never more.
When support allows
What else we run
We'd like all of these to be regular. Whether they happen depends on whether a partner, a campus or a venue makes them possible.
Study Jams
Group study through Kaggle Learn, books and free courses, with hands-on assignments and tutoring. The Kaggle track runs four weeks — machine learning and Pandas, then intermediate ML and feature engineering, then deep learning and computer vision, then time series and AI ethics. Finish a course and Kaggle issues you the certificate directly.
Monthly speaker sessions
Online talks, open to any member who wants to teach. Recorded, and the recording goes out on your channels too.
Tech exchange
Practitioners swapping what actually worked in production, and what didn't.
AI Safety
Sessions on alignment, evaluation and deploying models responsibly.
AI Governance
Policy and regulation, and what it means for people building here.
Hackathons
Campus and open build events, when a partner makes them possible.
ML Olympiad
A national Kaggle competition. We ran the 2024 edition as part of Google's worldwide ML Olympiad series.
Track record
What we've already run
Finished programs. Listed as history, not as a promise.
- CompletedFeb – Apr 2024
ML Olympiad 2024
A national Kaggle competition — the CO2 Emissions Prediction Challenge, predicting emissions per capita for 2030 from global development indicators. Hosted by Shahriar Azad Evan and Shuvro Pal under our former name, TFUG North Bengal, with Google Developer Group Dhaka. 300+ participants, and featured by Google on the Google Developers Blog as part of the worldwide ML Olympiad.
- CompletedMay – Aug 2026
Build With AI Hack Days
A one-off campus hackathon series, run while we had backing from MLH and Google for Developers. Seven stops — UITS, DIU, Rajshahi University, SEU, RUET and the EMK Center — free for every student who came.
- CompletedJuly 2026
Build With Gemma @Bangladesh
Our flagship hybrid hackathon at UIU, part of Google DeepMind's global Build with Gemma series. 82 teams from 42 universities; judged on Kaggle.