Start with DSA fundamentals — arrays, strings, linked lists, trees, graphs and dynamic programming — since most SDE interview rounds in India are DSA-heavy. Alongside this, build 1-2 solid resume projects and revise core CS subjects like OS, DBMS and OOPs. Once your basics are stable, add weekly mock interviews so you get comfortable explaining your approach out loud under time pressure. A consistent 3-4 month plan works well for most freshers.

Break it into phases: month one for core data structures (arrays, strings, hashmaps, linked lists, stacks, queues), month two for trees, graphs and recursion, and month three for dynamic programming plus full revision. Solve 2-3 problems daily instead of binge-solving on weekends, and maintain a sheet of patterns you keep getting wrong. In the final weeks, stop learning new topics and practise mixed problem sets under a timer, because interviews test pattern recognition, not memorisation.

Never go silent. Restate the problem, clarify the constraints, and think aloud — begin with a brute force solution, state its complexity, and then optimise step by step using hashmaps, sorting, two pointers or the relevant data structure. Interviewers evaluate your reasoning as much as your final code, so walking them through your approach, including your wrong turns, scores far better than jumping straight to an answer.

Most fresher interviews rotate around a predictable set: array problems like Kadane's algorithm, string questions such as anagrams and palindromes, linked list reversal and cycle detection, tree traversals, BFS/DFS on graphs, and entry-level dynamic programming. Instead of memorising answers, learn the underlying pattern behind each question type — interviewers tweak standard problems slightly to check whether you truly understood the concept.

Practice the way you will be tested: fix a daily slot, solve 2-3 problems on a timer, and speak your approach out loud before writing code, even when practising alone. Rotate topics weekly, redo failed problems after a few days, and once a week run a full 45-60 minute simulated round. Passive tutorial-watching does not build interview stamina — active, timed solving does.

Replicate real conditions — a 45-60 minute timer, a shared code editor and one person acting as the interviewer. Talk through your approach while coding, accept hints gracefully, and ask for feedback on communication, optimisation and code cleanliness afterwards. Peers and seniors are a good starting point, but a mock with an engineer who has actually cleared these rounds gives much sharper feedback. Expect medium-level mock coding interview questions from arrays, strings, trees and graphs in a typical fresher-level session, so cover those areas before your first mock.

A practical 3-4 month plan: month one for DSA fundamentals and 100-150 easy-to-medium problems; month two for advanced topics (trees, graphs, DP) plus 1-2 resume-worthy projects; month three for company-specific preparation, CS core subjects and resume polish; and the final weeks for mock interviews and revising your mistakes sheet. If you are starting from scratch or juggling college, stretch it to 5-6 months instead of rushing.

If you are disciplined, free resources plus consistent problem solving are genuinely enough to crack an SDE interview. A course helps when you need structure, deadlines, doubt-solving and curated problem sets instead of wandering across platforms. Many freshers in India take the middle path — self-study DSA, and invest only in 1:1 mentorship or mock interviews for personalised feedback, since that is where generic courses fall short.

Amazon's process usually involves online assessments followed by 2-3 interview rounds focused on DSA and Leadership Principles. Prepare DSA up to LeetCode medium level with a strong grip on arrays, trees, graphs and DP, and write 6-8 personal stories mapped to principles like Ownership, Customer Obsession and Bias for Action using the STAR format. Practise speaking your solutions aloud, because Amazon interviewers probe how you think, not just whether the code runs.

Get comfortable with Java's collections framework first — ArrayList, HashMap, HashSet, PriorityQueue and ArrayDeque — since choosing the right data structure is half the solution. Then solve topic-wise problems in Java itself on platforms like LeetCode and GeeksforGeeks rather than switching languages midway. Also revise Java-specific concepts such as String immutability, how HashMap works internally and the difference between == and .equals(), because these frequently come up alongside coding rounds.

GeeksforGeeks is excellent for company-wise archives and understanding what specific companies ask, but on its own it is usually not enough. Interviewers now twist standard problems, so you also need pattern-based practice on LeetCode or similar judges and a few mock interviews to test your live thinking. Treat GeeksforGeeks as one layer of your preparation — for company-specific questions and interview experiences — not your only resource.

Plenty of DSA interview questions PDF compilations circulate on coding communities, college groups and prep websites, and they are fine for last-minute revision. The catch is that solving from a PDF builds recognition, not recall — in a real interview nobody hands you options. A better habit is to maintain your own document of every problem you failed, along with the pattern used, and revise that before interviews.

It depends on your stage. Free peer mocks and community platforms are great for reducing initial nervousness, AI-based tools are handy for unlimited reps, and 1:1 mocks with experienced engineers from product companies give the most honest feedback on communication and approach. Whatever mock coding interview platform you pick, prioritise detailed human feedback over just a difficulty rating — that feedback is what actually improves your performance before placement season.

An AI mock coding interview is useful for high-volume practice — it is available anytime, does not make you nervous, and helps you get fluent at thinking aloud. What it cannot fully replicate is a human interviewer's follow-up questions, interruptions, hints and pressure, which is exactly what decides real interviews. A good mix is AI tools for daily reps and a few human mocks before important interviews.

A free mock coding interview — with a friend, a senior or a community member — is a great first step to get over stage fear and understand the format. The limitation is that free mocks often lack structured feedback on where you lost points, which is the entire purpose of a mock. Use free mocks for regular practice, and for companies you seriously care about, one or two paid 1:1 mocks with someone who has cleared those interviews is usually worth the investment.