Practice Platforms — Where You Grind Code¶
You cannot read your way to fluency. Somewhere between Month 2 and Month 12 you will solve 200-300 problems, review your solutions cold a week later, and only then does the syntax become reflex. This file tells you which platform to use for what, and in what order.
The Platforms, Ranked for Your Plan¶
Platform |
Free tier |
Paid tier |
Best for |
Verdict |
|---|---|---|---|---|
LeetCode |
2000+ free problems |
Premium: $35/mo or $159/yr (₹2,950 / ₹13,300) |
Big-tech study grind |
PRIMARY. Java 17/21 in browser. Premium worth it for M12 only. |
NeetCode.io + NeetCode Pro |
150 free “Blind 75/150” |
$149/yr (₹12,500) lifetime often on sale |
Structured DSA learning |
PRIMARY. Use |
Exercism (Java track) |
100% free |
N/A |
Reviewed idiomatic Java |
UNDERRATED. Real human mentors leave code review by PR. 130+ exercises. exercism.org/tracks/java |
HackerRank |
Most problems free |
$30-$60 for certifications |
study certifications for Indian recruiters |
SITUATIONAL. Some Indian firms (Infosys, Wipro, TCS) filter by HR-visible HackerRank certs. Complete the Java + Problem Solving cert. Skip otherwise. |
Codeforces |
Free |
Free |
Competitive programming |
SKIP. C++ dominates. Java in CF is a syntax handicap. |
AlgoExpert |
None (paid only) |
$99/yr (₹8,300) |
DSA video walkthroughs |
SKIP. NeetCode Pro is cheaper and better in 2026. |
CodeSignal |
Free tier |
$30/mo (₹2,500) |
Company-branded assessments |
SITUATIONAL. If a target company sends you a CodeSignal, practice it there. Otherwise skip. |
Codewars |
Free (community) |
$5-10/mo optional |
Small kata daily reps |
OPTIONAL. Fun, snackable. Good for 10-min lunch break drills in M1-M3. |
Project Euler |
Free |
N/A |
Math-heavy problems |
SKIP for studies. Fun on weekends only. |
Advent of Code |
Free (Dec each year) |
N/A |
December discipline |
DO IT. Every December. Use as annual Java re-warmup even after M13. adventofcode.com |
GeeksforGeeks Practice |
Free + ₹500-₹1500 courses |
₹5-15k Live courses |
India-flavored study problems |
CONSIDER. Some Indian company patterns (Infosys, Zoho puzzles) surface here first. |
The Order for Your 13 Months¶
Months 1-2: Syntax & Reflex¶
Exercism Java track — 30-40 exercises, request mentor review on every 3rd one. This is the fastest way back to Java fluency after 2 years off.
Codewars — 5-min kata daily. Optional.
Month 2 (dedicated DSA phase): Foundations¶
NeetCode Roadmap (free): Arrays → Two Pointers → Sliding Window → Stack → Binary Search → Linked List → Trees. Solve in Java 21. ~100 problems.
Don’t touch LeetCode Premium yet.
Months 3-6: Mixed reps¶
LeetCode Daily Challenge (30 min/day, free). Whatever Java the problem statement allows.
~5 problems/week during heads-down Java-language months.
Months 7-11: Reduce grind¶
2-3 problems/week only. Portfolio and system design take priority.
Month 12: study grind¶
LeetCode Premium ($35 for 1 month = ₹2,950): unlock company-tagged questions for target companies (Zoho, Freshworks, Razorpay, Swiggy, Amazon India, etc.).
NeetCode Pro if not already: complete “NeetCode 250” list.
HackerRank Java certification if applying to service companies.
Exercism revisit for mentored review of your solutions.
The Java-Specific Setup¶
Set LeetCode language to Java 17 or Java 21 (whichever is offered). Java 21 gives you records, pattern matching, sealed classes for cleaner solutions.
Learn the LeetCode Java stdlib shortcuts:
Arrays.sort,Collections.reverseOrder,PriorityQueue,Deque=ArrayDeque,HashMap.getOrDefault,List.of,Map.entry.Prefer
int[]overInteger[]when possible for perf on tight problems. Autoboxing kills LeetCode runtime scores.Common trap:
HashMap<int[], Integer>— array hash is identity-based, useList<Integer>as key or encode aslong/String.
The Anti-Pattern¶
Do not solve 500 problems in 2 months and think you’re ready. You will forget 90% of them. Space-repeat cold review — re-solve 5 old problems every Saturday without looking at your prior answer. This is the single practice most self-learners skip and most senior engineers rely on.
India Reality Check¶
Zoho studies historically test heavy CS fundamentals + puzzle-style problems. LeetCode Easy/Medium arrays, strings, DP is the right target. Zoho does NOT LeetCode-hard on most rounds — this is a difference from FAANG.
Freshworks, Razorpay, PhonePe, Swiggy, Zomato use LeetCode-style, harder end.
Service companies (Infosys, TCS, Wipro, Cognizant, Accenture) heavily use HackerRank + aptitude. HackerRank Java certification helps here more than LeetCode ranking.
Startup studies in Bangalore/Chennai increasingly do take-home + system design, less algorithmic. Reduce LeetCode weight accordingly.
Return to README.md · Next: 05_reference_docs.md