Data Structures and Algorithms Guide for Indian Interviews 2026

Not all DSA interviews are the same. FAANG India (Amazon, Google, Microsoft) tests medium-hard DSA across 4–5 rounds. Expect DP, graphs, trees, and heaps. Indian product unicorns (Flipkart, Swiggy, Razorpay, Meesho) test medium DSA, often with a practical twist: 'design a function' rather than a pure algorithm question. Service IT (TCS, Infosys, Wipro, Accenture) tests easy-medium arrays, strings, and sorting for fresher roles. Some service IT companies (Infosys InfyTQ, TCS NQT Advanced) have a distinct DSA section. Startups vary wildly: 0 to 2 DSA questions, often weighted toward system design and practical coding.

You need to own 8 core patterns to pass the majority of Indian product company interviews: (1) Two pointers: pair sum, remove duplicates, container with most water. (2) Sliding window: max subarray of size k, longest substring without repeating characters. (3) Binary search: not just sorted arrays, but binary search on the answer (minimum time to complete tasks). (4) Hash maps: two sum, frequency counting, grouping anagrams. (5) Stacks and queues: next greater element, valid parentheses, BFS. (6) Trees: inorder/preorder/postorder traversal, LCA, max depth, path sum. (7) Graphs: BFS, DFS, topological sort, union-find for connected components. (8) Dynamic programming: Fibonacci variants, 0/1 knapsack, longest common subsequence, coin change. Master these 8; everything else is a variation.

The honest answer, calibrated by tier: For service IT (TCS/Infy/Wipro) freshers: 50–75 Easy problems. Master arrays, strings, and sorting. For Indian product companies (Flipkart, Swiggy, Razorpay): 150–200 problems covering Easy + Medium, weighted toward the 8 patterns above. For FAANG India: 250–350 problems, including significant Medium and 30–40 Hard problems. The 'Blind 75' list is the best starting point: 75 problems that cover every important pattern. After the Blind 75, add the NeetCode 150 for breadth. After 150, the marginal return per problem drops significantly: depth of pattern understanding matters more than raw count.

Week 1–2: Arrays and strings (two pointers, sliding window, prefix sums). Do 15–20 problems per topic. Week 3: Hash maps and sets. Focus on frequency counting and lookup patterns. Week 4: Stacks, queues, and monotonic stacks. Week 5: Binary trees (recursion heavy: inorder, LCA, diameter, path sum). Week 6: Graphs (BFS, DFS, topological sort, union-find). Week 7: Dynamic programming (start with 1D DP: climbing stairs, house robber, coin change; then 2D: grid paths, LCS). Week 8: Mock interviews: 2 timed problems per day, treated as real interviews. Narrate your approach out loud. Review time complexity after every problem. This plan works for experienced engineers too: compress it to 4 weeks by doubling daily practice.

Every Indian product company interview expects you to state time and space complexity after each solution. The hierarchy: O(1) constant → O(log n) binary search → O(n) linear → O(n log n) sorting → O(n²) nested loops → O(2^n) exponential. Know these patterns by heart. A common trap: candidates solve a problem correctly but give wrong complexity. 'Merging two sorted arrays of size m and n' is O(m+n) time, not O(n log n). Always justify: 'The outer loop runs n times, the inner loop averages log n via binary search, so total is O(n log n).'

Mistake 1: Grinding LeetCode without understanding patterns. After 200 random problems, candidates still cannot solve new problems because they memorise solutions rather than patterns. Fix: solve a problem, identify the pattern, find 3 more problems that use the same pattern. Mistake 2: Skipping the brute force. Always start by stating the naive O(n²) or O(n) solution, then optimise. Interviewers credit the thought process, not just the optimal answer. Mistake 3: Silence. Indian candidates often solve in their heads before speaking. Western-style 'think aloud' interviews reward narration: say what you are thinking as you think it. Mistake 4: Not testing on edge cases. Always test: empty input, single element, all same elements, negative numbers, very large input.