The coding interview is the highest-stakes filter in Indian tech hiring: pass it and you get to the interesting rounds; fail it and the process ends. Yet most engineers prepare suboptimally: grinding 500 random LeetCode problems without a structured strategy, or practising in silence when interviews require explaining your thinking. This guide provides a realistic timeline, the specific DSA topics by priority, a structured 8-week plan, and how preparation differs for Indian product companies versus FAANG India.
Preparation time by target: Indian product companies Tier 1 (Razorpay, Swiggy, Flipkart, Zepto, Groww, PhonePe, CRED): 6-8 weeks of consistent daily practice (1-2 hours) if you have a solid CS foundation. Roughly 100-150 LeetCode problems. Global tech companies in India Tier 2 (Atlassian, LinkedIn, PayPal, SAP, Uber, Airbnb): 8-12 weeks. 150-250 problems, including system design practice. FAANG India (Google, Amazon, Meta, Microsoft, Apple): 3-6 months. 300+ problems, including hard-difficulty, plus system design and behavioural preparation. These timelines assume: you are a working engineer currently (not a full-time student), you can commit 1-2 hours daily on weekdays and 3-4 hours on weekends, and you have at least a basic data structures background (you remember what a BST is even if you cannot implement it from scratch). If you have not touched algorithms since your college data structures course, add 2 weeks at the start for concept review.
Topics by frequency in Indian product company interviews: Very high frequency (appear in 70%+ of rounds): arrays and strings (two-pointer technique: pair sum, container with most water, trapping rain water; sliding window: longest substring without repeating, max sum subarray of size K; prefix sums: range sum queries, subarray sum equals K), hash tables (two-sum variants, frequency counting, grouping anagrams), and trees (binary tree traversal: inorder/preorder/postorder iteratively and recursively; level-order BFS; LCA, Lowest Common Ancestor; validate BST; path sum problems). High frequency: dynamic programming (5 core patterns: 0/1 knapsack, unbounded knapsack, longest common subsequence, palindrome DP, interval DP, solve 4-5 examples of each pattern), graphs (BFS for shortest path in unweighted graph, DFS for connected components and cycle detection, topological sort for dependency ordering), and binary search on answer (search in rotated array, minimum in rotated sorted array, find first and last position). Medium frequency: heaps (k-th largest, merge K sorted lists, median from data stream), linked lists (reverse, detect cycle with Floyd's, merge sorted), and stack/queue (valid parentheses, monotonic stack for daily temperatures). Low priority for most Indian companies: segment trees, Fenwick trees, suffix arrays, Aho-Corasick (FAANG Google only).
Week 1-2 (Foundation, 30 easy problems): solve one problem per topic category to assess your baseline in each area. Focus on writing clean code and understanding time complexity. Read editorial solutions even when you solve correctly. Week 3-4 (Core patterns, 40 medium problems): arrays, strings, hash tables. Spend no more than 25 minutes per problem before looking at hints. After solving, write down the key insight in one sentence so you remember the pattern (not the specific problem). Week 5-6 (Trees and graphs, 40 medium problems): binary trees, BSTs, BFS, DFS. Include topological sort (Course Schedule I and II) and Union-Find (Number of Islands II, Redundant Connection). Week 7 (Dynamic programming, 20 medium problems): select exactly 5 DP patterns and solve 4 problems each. Quality of understanding over quantity here. Week 8 (Mock interviews, daily): simulate real interviews: timer set to 45 minutes, camera on, talk through your approach before coding. Use LeetCode mock mode, Pramp (free peer matching), or Interviewing.io (paid, with professional interviewers). The simulation in week 8 is more valuable than 100 more problems.
Indian product companies (Razorpay, Swiggy, Zepto, Groww): primarily medium-difficulty problems. Clean code and correct edge case handling are valued more than algorithmic cleverness. Interviewers give significant weight to how you communicate your approach: can you explain why you chose a hash map over a sorted array? Can you identify your own edge cases before the interviewer has to point them out? Hard problems appear occasionally at senior level. Practising aloud is essential. Google India: known for the hardest coding bar among FAANG India. Hard graph problems (shortest path with constraints, Euler circuit), tricky DP (interval scheduling, DP on trees), and problems requiring number theory or combinatorics insight appear. 300+ problems recommended, including 50+ hard problems. Amazon India: medium problems primarily. Emphasis on OOP design and arrays. Amazon's Leadership Principles are weighted heavily: the behavioural round matters as much as the coding round. Microsoft India: gentler than Google India. Medium problems, emphasis on code quality and discussing alternatives. A medium-difficulty engineer who communicates clearly often outperforms a strong algorithmist who cannot explain their thinking.