I was doing technical interview prep wrong for years. Here's the best advice from Senior FAANG SWEs.

I used to spend 1 hour trying to solve a problem. Then, 15 minutes learning the solution.

That's backwards. And it cost me.

I'm a FAANG SWE and CS instructor. I failed a mock interview with a Senior FAANG Engineer after years of giving interview advice. That failure forced me to rebuild how I study. Now that same engineer is mentoring me into the top 0.5% of technical interviewing.

Here's the study system I use now.

My Biggest Mistake

Interviewers expect the optimal solution. Not a working solution. The optimal one.

So grinding for an hour on your own approach teaches you the wrong thing. You're reinforcing your way. But your way wasn't the right solution. That's why you were stuck.

The learning happens when you study the best solution and understand why your approach fell short.

Flip your time. Spend less time solving. Spend more time understanding.

How to Practice

Here's the time budget I follow:

• Spend 5 minutes trying to solve an Easy problem.

• Spend 10 minutes trying to solve a Medium problem.

• Then spend the same amount of time, plus or minus 10 to 20 minutes, understanding the solution. Keep learnin until you understand.

  
  

Don't waste time. Sometimes you can get stuck and work on something for far too long without progress. Recognize when that's happening and move to the solution.

To understand a problem you're stuck on:

• Watch videos

• Read existing solutions

• Use an LLM to explain and break down the solution for you

• If you think you're close, paste your code into a chat tool to find the specific issue

  
  

For every problem, you should know 3 solutions: the brute force, the optimal, and the most optimal. Know why each one exists. Optimize for runtime complexity every time.

Learn the repeatable patterns. That's the whole game.

How Many Problems to Solve

There is not a set number, but often 200 to 400 gives good results in the current market.

If that's not realistic, the Blind 75 to 150 is a strong starting point.

But the number is not the goal. I met a candidate who solved 500 LeetCode problems and failed Amazon interviews twice. He did not understand the underlying patterns, data structures, or algorithms.

Another candidate solved the problem correctly at a FAANG interview. Still no offer. She knew the material but rushed the process because she never mock-interviewed with a real person.

You must do both. Understand deeply. And practice live.

The Bar Is Higher, Not the Questions

The problems haven't gotten harder. But there's more competition now. Problems are expected to be solved with little to no mistakes.

Less room for error than there used to be. That's the real change.

Company-Specific Preparation

Where you can, do problems related to the specific company you're interviewing with.

• Solve the most recent 60 questions asked at that company within the last 30 to 90 days.

• This works especially well for Big Tech.

• For most Big Tech companies, having solved that exact problem or one very close to it makes the difference. You can find these on LeetCode.

Test Your Code in the Interview

Don't write and submit. Walk through your code. Use print statements, assertions, or exceptions after you finish.

Treat it like real code. Interviewers notice.

Use LeetCode the Right Way

LeetCode is a great tool when used properly.

Start with an easy problem. LeetCode will surface similar questions, some with the same difficulty, some harder. Solve the related problems in sequence.

That's how you deepen the pattern, not just memorize one problem.

Structure Your Study Sessions

• Study in blocks of 1 to 4 hours. 2 hours is often the sweet spot.

• Time-box yourself. Don't let a session run unstructured.

• Repetition matters. Come back to problems the next day, or a week later, or two.

• Aim to be genuinely strong across all topics, not just your favorites. In today's interviews, thoroughness matters.

  
  

Know the core concepts cold.

Organize Your Study

Create a spreadsheet of every question you've solved.

• Track the problems you understand well.

• Track the problems you need to brush up on.

Review the one you need to brush up on first. Simple system. It works.

Summary

• Spend more time learning solutions than grinding on problems. Interviewers expect the optimal solution.

• Know the brute force, optimal, and most optimal for every problem.

• 200+ problems great. Blind 75 to 150 is a strong start. Understanding beats volume.

• Do company-specific prep: the last 60 questions from the last 30 to 90 days.

• Track everything in a spreadsheet.

  
  

I made these mistakes with a FAANG job on my resume. You don't have to.

Subscribe to our newsletter and get updated each time we release an article.