A free, certification course on GenAI × LLM mastery.
Unlock the technology behind intelligent systems: NLP, large language models, transformers, RAG and agents. Three weeks of structured learning with weekly assignments and industry expert sessions, then one week to build a GenAI capstone.
Built on the same idea as Summer Analytics, the club's long-running machine learning course: a little every day beats a lot once.
There is no shortage of material on language models; there is a shortage of order. Most people who start stop at the videos. This course is arranged so that the thing you build each day is small enough to finish, and the thing you build each weekend is worth keeping.
Rather than a tour of topics, it follows a single sentence through a model — tokenised, embedded, attended over, decoded — and then treats retrieval, fine-tuning and agents as the ways of changing what comes out.
NLP foundations, large language models, transformers, RAG, AI agents and multi-agent systems — from understanding how they work to building with them.
The Consulting & Analytics Club at IIT Guwahati is the student group behind Summer Analytics, which in 2025 drew more than 23,000 registrations from over 80 countries. This course is run by the club's machine learning team, with content curated from public lectures and notebooks; the assignments and the capstone are ours.
Weeks are released on Mondays. Resources for a released week are on the course dashboard.
Why text is hard for neural networks, what attention fixed, and what a transformer block is actually doing.
embeddingssequence modelsself-attentionthe transformer block
The lifecycle from pretraining to instruction tuning, then the practical half: decoding, prompting, local models, evaluation.
training lifecycledecodingpromptinglocal models + evaluation
The ways to make a general model work for you. Retrieval changes what it knows; fine-tuning changes how it behaves; agents let it act.
embeddings & similarityRAGfine-tuning with LoRAagents & multi-agent systems
No new material. One problem statement, released on the Monday, and a week to build against it and submit.
problem statementbuildsubmitshowcase
Weeks are released on Mondays. Assignment deadlines are Sundays, 11:59 PM IST.
| 2026-09-28 | Week 1 released |
| 2026-10-05 | Week 2 released |
| 2026-10-12 | Week 3 released |
| 2026-10-19 | Week 4 released |
| 2026-10-26 | Final submissions close |
Anyone who wants to understand what powers modern GenAI, whether you are starting out or looking to go deeper into LLMs. No prior NLP or LLM knowledge is required.
Being comfortable in Python and having seen a machine learning model before will make the hands-on tasks smoother, but neither is a hard requirement.
Nothing needs to be installed locally until the second week. No credit card is required for any of it.
Free to participate · open to everyone · no prior NLP or LLM knowledge required.
