GATE DA Syllabus 2027: Exam Pattern, Weightage & Preparation

AI is moving fast, and the research behind it isn't slowing down. GATE DA is one of the cleanest ways into that world, and the score opens up M.Tech, research and PhD routes at the IITs and other strong institutes.

It also sets you up for the right kind of work. Data science and AI is a research field at its core, not just a set of tools you learn to operate. The syllabus covers what sits underneath the libraries, so you can build methods of your own instead of only running what other people built.

If you've looked at any GATE paper before, this one will feel familiar. You get 65 questions for 100 marks in three hours, all on computer. The marks are split between General Aptitude and the core subject, and every question is either an MCQ, an MSQ or NAT type. The one thing that's different, and the only thing worth planning around, is how those marks get spread across the topics.

How the question mix has moved

Over the three papers we've had so far, the balance has shifted a little. MCQs have eased down, and NAT questions (the ones where you type the answer, with no options to work backwards from) climbed to their biggest share yet in 2026.

GATE never publishes a fixed subject-wise split, and it moves around from year to year. The chart below takes the average across all three papers so far.

One caution: three papers really isn't much data. It's enough to say what has happened so far, but not enough to promise it'll keep happening — take the shifts below as planning hints, not hard rules, and weight the most recent paper more than the first one.

What's moved, and which way

Here's the clearest shift: DBMS roughly tripled from 6 marks to 18, while Calculus & Optimization fell to just 2 marks in 2026.

You'll see some coaching sites put Machine Learning or Programming & DSA at the top, usually because they're going off the earlier papers. Once you add 2026 in, our three-year read puts Probability & Statistics first — but none of this is officially fixed, so treat any ranking (ours included) as a guide, not a promise.

IIT Madras hasn't changed the DA syllabus for 2027, so this all still holds. Here it is as a quick reference, enough to plan around without reading like a textbook.

The weightage tells you what to study. These four are more about how you study it, so that it stays with you.

01 · Keep the whole subject in rolling revision

Don't treat modules as a checklist you finish and forget. As you move forward, keep going back: before Module 2, redo Module 1; before Module 3, redo 1 and 2, and so on. And revision means solving questions without looking at the answers, old tests and quizzes included, not just re-reading.

02 · Go over every test and split the losses

After every test, split the marks you lost into two piles. The ones you knew but still got wrong are an accuracy problem: stop giving away marks you'd already earned. The ones you didn't know are a gap: go and learn the topic. Same lost mark, two different fixes.

03 · Read the whole question, not just the keywords

The most common way to lose easy marks is misreading the question. Plenty of strong students skim the keywords and answer the question they assumed was asked. Slow down, read it fully, and you'll clear far more of the paper. It's the cheapest accuracy there is.

04 · Pick one path and stop hoarding resources

More material isn't more preparation. Bouncing between ten question banks feels productive but starves the thing that works: going over the same structured material again and again. The maths and the ML algorithms only click when you keep returning to them. Pick one course or resource set and go deep instead of wide.

Double-check all of this on the official IIT Madras site, since these are the bits that change every year.

Sources: official GATE DA syllabus (IIT Madras) and an in-house question-by-question analysis of the 2024, 2025 & 2026 papers. Weightage is indicative, not an official fixed distribution.