I got Google MLE (L4) by practicing 200 LC questions(East Coast, 2021)

In this interview series, I summarize interview experiences from people I helped. I hope it helps you in preparing for Google MLE interview.

SC has background in Computer Science with 4 years of experience. He got Google Software Engineer, Machine Learning position (L4) in East Coast. He wants to share his experience for others.

Discussion

1. Do you focus on companies tagged specifically?

Only the week before an interview for that company. The frequency tags were fairly accurate for Amazon, Facebook, and Two Sigma but Google used fairly unique problems I haven’t seen on Leetcode before.

2. How many questions did you solve? How many weeks/hours you spent on LC practice.

All said and done about 200 questions on Leetcode. I feel I over prepared. I focused my efforts on mostly Medium problems and occasionally would “blitz” 5 easy questions in 30 minutes. Spent some time on hards but not much. I did probably 10 hours a week for 2–3 months. Just estimating hard to say. Probably did something around 50 Leetcode questions during actual interviews since I spoke with a lot of companies.

Book mock interview.

3. If you start again what is the one thing you want to do differently in terms of preparation?

I think I did a good job of preparing and I wouldn’t change anything.

4. What other tips would you recommend?

Learn all of the fundamental data structures and algorithms so that you understand them fully. Do NOT learn DS&A specifically in terms of solving Leetcode problems.

Solve problems based on different patterns. Then when you get a problem in an interview identify what pattern it is. For example, understand Two Pointer patterns and recognize palindrome problems as frequently using Two Pointers.

Keep note of small tips that can come up for improving time complexity. Obvious example — caching with hashmaps / dictionaries. Less obvious — anagrams can be solved NLogN with sorting. Or you can do in O(N) by caching with a counter using multiplying of primes or storing with an array of counts.

5. Do you find mock interviews helpful? In terms of ML knowledge, which books/resources you find helpful.

Doing mock interviews if you have never interviewed before would definitely beneficial, but I have done so many Leetcode style interviews now that I feel like I have it down pat and wouldn’t invest in it now.

Support me to read more content like this. If you have any questions, you can drop me an email: [email protected].

Read more

• https://github.com/khangich/machine-learning-interview
• https://www.educative.io/courses/machine-learning-system-design
• https://www.interviewquery.com/courses/ml-system-design