How to Crack DSA for Placements in 90 Days

Ninety days is enough time to go from "I know loops" to "interview-ready", but only if you spend it well. Most students fail not from lack of effort but from bad allocation: Six weeks perfecting arrays, then panicking through graphs and DP in the final fortnight. This plan front-loads the hard topics while you still have runway.

The ground rules

• 3–4 problems a day, every day beats 20 problems one day and zero for a week. Consistency compounds; cramming doesn't.

• Revision is non-negotiable. A solved problem you can't re-derive in 30 days is an unsolved problem.

• Track honestly. "Looked at the solution" is not "solved." Mark it for revision instead.

Phase 1: Foundations (Days 1–21)

Arrays, strings, hashing, two pointers, sliding window, basic recursion. These are the alphabet of every harder topic. Aim to recognise the pattern, not memorise solutions.

Milestone: ~70 problems, and you can spot a sliding-window problem on sight.

Phase 2: Core structures (Days 22–50)

Linked lists, stacks & queues, trees (traversals, BST), heaps. This is where most people slow down, so budget for it. Do trees thoroughly; they're the gateway to graphs.

Milestone: ~90 problems cumulative; comfortable with recursion on trees.

Phase 3: The hard topics (Days 51–75)

Graphs (BFS/DFS, topological sort, shortest paths) and dynamic programming. This is where interviews are won and lost. If you've front-loaded correctly, you have three full weeks of fresh energy for the material everyone else rushes.

Milestone: You can model a problem as a graph and choose the right traversal; you recognise the 6–7 canonical DP patterns.

Phase 4: Revision & mocks (Days 76–90)

No new topics. Re-solve everything you flagged, do timed mocks, and practise talking through your approach out loud. The goal of the last two weeks is speed and composure, not coverage.

Milestone: Two clean mock interviews where you narrate your reasoning end-to-end.

Why role-based ordering matters here

This timeline assumes a generalist SDE target. If you're aiming at a specific role, the weights shift; a frontend or data role lets you compress Phase 3. That's the advantage of letting your role set the order instead of marching through a sheet top to bottom. Set your role, let the plan reshape, and track it without the spreadsheet.

Milestone problems for each phase

A phase is done when its milestone problems are solvable from a blank editor, not when the day count runs out. Phase 1, foundations: Two Sum, Valid Anagram, Best Time to Buy and Sell Stock, Longest Substring Without Repeating Characters and 3Sum. Phase 2, core structures: Reverse Linked List, Linked List Cycle, Valid Parentheses, Min Stack, Binary Tree Level Order Traversal, Kth Largest Element in an Array. Phase 3, the hard topics: Number of Islands as the entry to graphs, then the Graph topic hub for topological sort and shortest paths; Climbing Stairs, Coin Change and Word Break as the three dynamic programming shapes that recur. Phase 4 has no new problems by design; it re-solves the ones above under a timer. If a milestone problem still needs the note on day 76, that topic, not the calendar, decides what the last two weeks are spent on.

Related on DSA Tracker

• Off-campus placement plan
• 7-day revision plan
• Dynamic Programming topic hub
• Graph topic hub
• Consensus 75

Further reading: Binary search trees on Wikipedia.