GATE DA 2026 Question Paper (PYQ) with Subject-Wise Analysis & Weightage | Piyush Wairale
By Piyush Wairale · IIT Madras

GATE DA 2026 Question Paper — Subject-Wise Analysis & Weightage

A complete breakdown of the GATE 2026 Data Science & Artificial Intelligence (DA) previous year paper — marks distribution, important topics, difficulty, and a question-by-question map of all 65 questions. Conducted by IIT Guwahati on 15 February 2026.

65
Questions
100
Marks
3 hrs
Duration
8
Subjects
TL;DR — Key Takeaways
  • DBMS was the single highest-weighted subject with 18 marks, driven heavily by B+ Trees, SQL, relational algebra and normalization.
  • PDSA (16) and Machine Learning (16) tied just behind — both application-heavy, so expect to trace code and compute metrics, not just recall theory.
  • DBMS + PDSA + ML = 50 marks — half the paper comes from databases, coding and machine learning.
  • The math core was spread across Probability & Statistics (11), Calculus & Optimization (9) and Linear Algebra (8) — about 28 marks combined.
  • General Aptitude held its usual 15 marks; Artificial Intelligence rounded things out at 8.
01 · Exam Snapshot

GATE DA 2026 exam pattern

GATE DA 2026 was a 3-hour Computer-Based Test with 65 questions for 100 marks, split into a common General Aptitude (GA) section worth 15 marks and a core Data Science & AI section worth 85 marks. The paper mixed MCQ, MSQ (multiple-select) and NAT (numerical answer type) questions.

Section Questions Marks Marking
General Aptitude (Q1–Q10) 10 15 5 × 1M + 5 × 2M
Core DA — 1 mark (Q11–Q35) 25 25 MCQ / MSQ / NAT
Core DA — 2 marks (Q36–Q65) 30 60 MCQ / MSQ / NAT
Total 65 100

Negative marking applies to MCQs only: −1/3 for a 1-mark MCQ and −2/3 for a 2-mark MCQ. MSQ and NAT questions carry no negative marking.

02 · The Big Picture

GATE DA 2026 subject-wise weightage

Here is the official subject-wise distribution of all 100 marks, split by 1-mark and 2-mark questions. The bars below are scaled to the highest-weighted subject so you can see priorities at a glance.

Subject 1-mark Qs 2-mark Qs Total marks
General Aptitude5515
Calculus & Optimization339
Probability & Statistics4311
Linear Algebra238
DBMS & Data Warehousing4718
Programming, DS & Algorithms (PDSA)4616
Machine Learning4616
Artificial Intelligence428
Total3035100
DBMS & Data Warehousing 18 marks
Programming, DS & Algorithms 16 marks
Machine Learning 16 marks
General Aptitude 15 marks
Probability & Statistics 11 marks
Calculus & Optimization 9 marks
Linear Algebra 8 marks
Artificial Intelligence 8 marks

Note: weightage follows Piyush Wairale's official subject classification. Optimization-based questions (gradient descent, regularization objective, series) are grouped under Calculus & Optimization. The question-by-question table further below is an indicative topic reference — a few boundary questions may be grouped differently there than in this headline count.

▶ Video Walkthrough

Watch the full GATE DA 2026 paper analysis

Piyush Wairale walks through the GATE DA 2026 question paper, subject-wise weightage, and the topics that mattered most — in this detailed video breakdown.

03 · Deep Dive

Important topics, subject by subject

Probability & Statistics

11 marks

A high-value math area where questions leaned on distribution properties and clean conceptual reasoning rather than heavy calculation.

Must-know topics: Normal, t, Chi-square, Exponential (memoryless property), Bernoulli, Poisson & Uniform distributions; Bayes theorem; CDF properties; variance, covariance & correlation; counting/combinatorics.
Representative questions: Q19, Q20, Q28, Q33, Q34, Q44, Q45, Q53, Q54, Q57, Q62, Q63, Q64.

Database Management & Warehousing

18 marks

The highest-weighted subject in the paper (18 marks) — and one that rewards hands-on query practice. Two separate B+ Tree questions plus deep SQL made this a make-or-break area.

Must-know topics: B+ Trees (order, inserts, node pointers), SQL (correlated subqueries, GROUP BY, aggregates), relational algebra & tuple relational calculus, functional dependencies & candidate keys, normalization (3NF), ER modeling, OLAP (roll-up, drill-down, cuboids).
Representative questions: Q17, Q18, Q26, Q32, Q41, Q42, Q43, Q51, Q59, Q60, Q61.

Programming, Data Structures & Algorithms

16 marks

Heavily application-based. Expect to trace Python code and reason about algorithm behaviour rather than recall definitions.

Must-know topics: Sorting (quicksort recurrence, bubble vs. insertion sort), binary search, binary tree traversals, graph BFS/DFS & reachability, recursion & the runtime stack, Python gotchas (mutable default arguments, closures).
Representative questions: Q15, Q16, Q25, Q31, Q39, Q40, Q49, Q50, Q58.

Machine Learning

16 marks

Numerical and definitional in equal measure. Several questions asked you to compute a metric or count parameters.

Must-know topics: PCA, cross-validation (LOOCV), clustering (hierarchical, K-medoid), regression & regularization (Ridge / L1 vs L2), classification metrics (accuracy, precision, recall), gradient descent (SGD), MLP parameter counting, bias–variance tradeoff.
Representative questions: Q11, Q12, Q23, Q29, Q36, Q37, Q47, Q55, Q56.

Linear Algebra

8 marks

A reliable, concept-driven scorer. The projection matrix M = I − (1/n)11ᵀ appeared twice (Q52 & Q65).

Must-know topics: eigenvalues, rotation matrices & matrix powers, projection & idempotent matrices, quadratic forms (max xᵀAx), vector spaces & subspaces.
Representative questions: Q21, Q22, Q46, Q52, Q65.

Artificial Intelligence

8 marks

Logic-heavy this year. First-order logic showed up in multiple questions, so precision with quantifiers pays off.

Must-know topics: informed vs. uninformed search (A*, BFS, DFS), first-order / predicate logic & validity, logical entailment & equivalence, minimax & game trees.
Representative questions: Q13, Q14, Q24, Q30, Q38, Q48.

Calculus & Optimization

9 marks

Bigger than it looks once optimization is counted here. Gradient descent, regularization objectives and series questions all live in this bucket — a genuinely high-yield area, not an afterthought.

Must-know topics: maxima/minima & roots of functions, gradient descent / stochastic gradient descent, regularization (ridge) objectives, sequences & series (geometric series summation), convergence.
Representative questions: Q27, Q29, Q35 (plus optimization steps embedded in Q37 & Q55).

04 · Full Map

Question-by-question breakdown

Every one of the 65 questions mapped to its subject and topic. Use this to spot exactly which concepts were tested.

Q Marks Subject Topic tested
11GAVerbal — analogy
21GANumerical — digit product/sum
31GASpatial reasoning — puzzle
41GAQuantitative — logarithms
51GALogical reasoning — conditionals
62GAVerbal — prepositions
72GANumerical — mean/median/mode
82GAAnalytical reasoning — logic puzzle
92GANumerical — number puzzle
102GAGeometry — circle area
111MLPCA — principal component angles
121MLLOOCV validation splits
131AIUninformed vs informed search (A*)
141AIFirst-order logic statements
151PDSAQuicksort recurrence relation
161PDSAPython mutable default argument
171DBMSFunctional dependencies, candidate keys
181DWOLAP — drill-down
191Prob & StatsProbability — subset product parity
201Prob & StatsInteger solutions probability
211Linear AlgebraRotation matrix powers (M²⁰²⁶)
221Linear AlgebraDisk ∩ subspace area
231MLTask–algorithm matching
241AILogical entailment
251PDSABinary tree traversals
261DBMSForeign key constraint
271CalculusRoots, maxima & minima
281Prob & StatsNormal vs t-distribution CDF
291CalculusSGD parameter update (optimization)
301AIMinimax game tree
311PDSABinary search — max comparisons
321DBMSB+ Tree — node pointers
331Prob & StatsCounting — bijections f(f(n))=n
341Prob & StatsExponential — memoryless property
351CalculusDouble summation — geometric series
362MLHierarchical clustering (Manhattan)
372MLRidge regression properties
382AIPropositional logic equivalence
392PDSARecursion — call count (Fibonacci-like)
402PDSADirected graph BFS/DFS reachability
412DBMSB+ Tree insertion — root split
422DBMSRelational algebra evaluation
432DWData cuboids from concept hierarchies
442Prob & StatsVariance of (2X−1)Y
452Prob & StatsPoisson CDF limit
462Linear AlgebraEigenvalues of rotation matrix
472MLAccuracy / precision / recall
482AIFirst-order logic validity
492PDSABubble sort vs insertion sort
502PDSAPython closures & scope
512DBMSSQL — team size (GROUP BY)
522Linear AlgebraProjection / centering matrix
532Prob & StatsChi-square distributions
542Prob & StatsCDF properties
552MLRidge regression loss (MAE)
562MLMLP — learnable parameters
572Prob & StatsBayes theorem — disease test
582PDSARecursive bubble sort — swap count
592DBMSTuple relational calculus
602DBMSSQL — correlated subquery
612DBMSER model → 3NF relations
622Prob & StatsVariance from pairwise sum
632Prob & StatsCorrelation of X, Y
642Prob & StatsBernoulli matrix — row/col sums
652Linear AlgebraQuadratic form max xᵀAx
05 · What This Means For You

How to prepare, based on 2026 trends

  1. Make DBMS your #1 scoring machine. At 18 marks it is the highest-return subject. B+ Trees and SQL are learnable to near-100% accuracy with drilled practice — don't leave marks on the table here.
  2. Practise code tracing daily for PDSA (16). Questions demanded you predict Python output (closures, mutable defaults) and count recursive calls. Reading theory is not enough — trace by hand.
  3. Make ML numerical, not just conceptual (16). Be able to compute precision/recall, count MLP parameters, and run an SGD step. Definitions alone won't clear these.
  4. Lock in General Aptitude (15). A guaranteed 15 marks with steady practice — the cheapest marks in the paper.
  5. Build the math core: Prob & Stats (11) + Calculus & Optimization (9) + Linear Algebra (8). Master distributions and Bayes theorem, gradient descent and regularization, and projection/eigenvalue problems — together nearly 28 marks.
  6. Keep AI (8) tight. First-order logic and search are compact, high-accuracy topics — quick to revise, easy to bank.
By Piyush Wairale · IIT Madras

Start your GATE DA 2027 prep today

Everything you need to crack GATE Data Science & AI 2027 — from daily concept practice to full-length mocks with detailed solutions.

Popular

GATE DA Daily Practice Series (DPS)

5 carefully selected GATE DA questions delivered every day. Build exam-day speed and deepen concept clarity over 90 days.

₹699 ₹1,500
Enroll Now →
2027 Batch

GATE DA 2027 Test Series — Full Mock Tests

Comprehensive GATE Data Science & AI mock tests covering the full syllabus. Timed exams, detailed solutions, and All India Rank to track your progress.

₹1,199 ₹1,500 20% OFF
Enroll Now →
⭐ Bestseller · Best Value

Best GATE DA 2027 — Complete Course + Test Series + DPS Bundle

India's top-ranked GATE Data Science & AI course — full syllabus, 250+ hrs of video lectures, AI-powered mock tests, and 5000+ practice questions. Everything in one bundle at the best price.

₹6,000 ₹13,000 60% OFF
Enroll Now →
Instructor: Piyush Wairale · IIT Madras · English
06 · FAQ

GATE DA 2026 paper — frequently asked questions

Which subject had the highest weightage in GATE DA 2026?

DBMS & Data Warehousing had the highest weightage with 18 marks. PDSA and Machine Learning followed with 16 marks each.

How many questions were asked in GATE DA 2026?

There were 65 questions for 100 marks: 10 General Aptitude questions (15 marks) and 55 core Data Science & AI questions (85 marks). Q1–Q5 and Q11–Q35 carried 1 mark each; Q6–Q10 and Q36–Q65 carried 2 marks each.

What is the full subject-wise weightage of GATE DA 2026?

DBMS & Data Warehousing 18, PDSA 16, Machine Learning 16, General Aptitude 15, Probability & Statistics 11, Calculus & Optimization 9, Linear Algebra 8, and Artificial Intelligence 8 — totalling 100 marks.

Was GATE DA 2026 difficult?

Overall moderate to moderately difficult. Programming and DBMS were application-heavy (Python closures, correlated SQL subqueries, B+ Tree inserts), while Probability & Statistics rewarded conceptual clarity over rote formulas.

Where can I get GATE DA 2026 solutions and more PYQs?

Detailed solutions, topic-wise practice, and full-length mock tests for GATE DA are available in Piyush Wairale's GATE DA 2027 course and test series.

PW
Piyush Wairale
M.Tech, IIT Madras · Ex-Instructor, IIT Madras BS Programme · Founder, PiyushAI Edtech. Mentoring 20,000+ aspirants across GATE DA, Bank IT Officer & PSU exams.

Analysis based on the official GATE 2026 Data Science & AI question paper (IIT Guwahati). © PiyushAI Edtech. Subject classification reflects the official GATE DA syllabus; boundary questions are noted where relevant.