Taro Logo

Machine Learning Engineer Interview Experience - San Francisco, California

February 1, 2021
Positive ExperienceNo Offer

Process

A recruiter reached out to me in January (we actually started communicating back in June 2020).

Phone Screen Had a phone screen in early February. See post in the description.

Full Loop The full loop was spread over two separate days (two days apart).

  • Day 1
  • Day 2

Decision was made on the second day after the interview.

Questions

I summarized my experience in the referred post above.

Scheduled a full loop two months ahead to have ample time for prep (but couldn't really cover much).

First coding interview: Two questions

a. Find the maximum sub-tree sum. I couldn't find a related question on LeetCode. Essentially, consider every node in the tree as the root, calculate the sum of the nodes, and then return the maximum such sum. b. Variation of https://leetcode.com/problems/word-break/. Return the output as a concatenation with the words from the dictionary separated by a single space. E.g., string = "catsanddogs", wordDict = ["cat", "sand", "cats", "dog", "and"], an output = "cat sand dogs" or "cats and dogs".

Second coding interview: Two questions

a. Return the top k integers with the highest frequencies in an array: https://leetcode.com/problems/top-k-frequent-elements/ b. Course schedule: https://leetcode.com/problems/course-schedule-ii/

Behavioral interview + one coding question

a. Typical workplace behavior questions. b. Find the length of the longest sub-array whose sum is target.

System design: Design a Machine Learning app that makes recommendations to users for places to visit along their trip.

Focus was on the ML pipeline: Data, Features, Evaluation, Model-building, Feedback, Online testing, Offline testing.

Personal Assessment: Didn't give a good account of myself (first time system design).

ML system design: Design an E2E classification pipeline for Facebook Marketplace.

Users post visual (photos) + textual descriptions. Lots of focus on the Model training practice + Feature Engineering + Feedback loop, as well as emphasis on theory.

Personal Assessment: Way better than round 4.

Calibration (Behavior + Coding):

a. Behavior similar to round 3. b. Given a linked list L, a value val, and position pos, insert val at the pos-th position in L (not replace).

Was this helpful?

Interview Statistics

The following metrics were computed from 11 interview experiences for the Meta Machine Learning Engineer role in San Francisco, California.

Success Rate

0%
Pass Rate

Meta's interview process for their Machine Learning Engineer roles in San Francisco, California is extremely selective, failing the vast majority of engineers.

Experience Rating

Positive73%
Neutral9%
Negative18%

Candidates reported having very good feelings for Meta's Machine Learning Engineer interview process in San Francisco, California.

Meta Work Experiences