How Many LeetCode Problems Should I Do Per Day?

Do 1 to 2 LeetCode problems a day if you have one hour, 2 to 3 if you have two hours, and 3 to 4 if you have three hours. Over a preparation period of two to three months that comes to 150 to 300 problems. That is enough for interviews at Google, Amazon, Meta and similar companies when the problems are chosen by pattern rather than at random. The number per day matters less than two other things: that you do it every day, and that each problem teaches a pattern you can reuse.

A Daily Plan by Available Hours

Medium is the level to spend most time on. It is the difficulty most interview questions are set at, and a hard problem is usually two mediums combined. Budget about 30 to 45 minutes per medium: 20 to 25 minutes to attempt it, then the rest to read the best solution and write it again yourself.

Do not go past four hours a day. After that, retention drops and you start matching on the answer instead of the reasoning. Five problems done carefully are worth more than ten done quickly.

Is One LeetCode a Day Enough?

One problem a day is enough if you continue for three months or more and choose the problems by pattern. That is about 90 problems, which covers the 20 or so core patterns once each with a few repeats. It is not enough in the two weeks before an interview, where 3 to 4 a day plus timed practice is the realistic minimum.

For a student in the second or third year, one a day is the right pace. The goal at that stage is to learn the data structures properly, not to reach a count. Raise the pace to 2 to 3 a day in the two months before placement or internship interviews.

How Many Problems for FAANG or Google

Plan on 150 to 300 problems, spread over 15 to 20 patterns, before an interview at a large tech company. People who pass with 100 have usually chosen those 100 well. People who fail with 500 have usually solved them in isolation and cannot recognize the pattern in a new problem. Company-specific lists matter less than pattern coverage, because interviewers rotate their questions.

For what these companies actually ask, see does Amazon ask LeetCode questions and the top coding interview questions at Google.

How Many LeetCode Problems Are There

LeetCode has more than 3,000 problems, and the count grows every week. Nobody solves them all, and nobody needs to. About a quarter are easy, half are medium, and a quarter are hard. The number you need is the number that covers the patterns, which is why a curated list of 75 to 150 problems is the standard starting point.

The Pattern-Based Approach

A pattern is a reusable solution structure: two pointers, sliding window, fast and slow pointers, merge intervals, binary search on the answer, BFS and DFS on trees and graphs, top K with a heap, backtracking, dynamic programming. Learning by pattern means solving five to ten problems of one pattern in a row until the structure is automatic, then moving to the next. When a new problem appears, you match it to a pattern in the first two minutes instead of starting from nothing.

Here is a working weekly routine on one hour a day.

• Monday to Thursday. One new medium from the current pattern, 45 minutes, then 15 minutes writing the solution again without looking.
• Friday. One problem from an earlier pattern, chosen at random, to check that it stayed learned.
• Saturday. One hard problem that combines two patterns, or a timed 45-minute mock question.
• Sunday. Re-solve the two problems from the week you found hardest. No new problems.

Tracking What Counts

Count patterns learned, not problems solved. Keep a list with three columns: problem, pattern, and whether you solved it without help. A problem you needed the editorial for goes back on the list for a re-solve in one week. When a pattern has five problems solved without help, it is done, and the daily number for that pattern can drop to one review a week.

How to Prepare

• Follow a pattern order rather than the problem list. Grokking the Coding Interview groups problems by pattern, so each day's problem uses what the previous one taught.
• Short on time? Start with a curated set. Grokking 75 Top Coding Interview Questions is sized for one to two problems a day over about two months.
• Repair the data structures first if mediums feel impossible. Grokking Data Structures for Coding Interviews covers arrays, hash maps, trees and heaps with exercises.
• Add timed practice in the last two weeks. A mock interview shows whether you can solve a medium in 25 minutes while explaining it.