Google PhD New Grad AI/ML Loop: Mahjong Hand and Prefix Search Onsite

Role: Software Engineer, PhD Early Career — AI/ML
Result: Entered Team Match
Report time: March 2026
Onsite location: Bay Area

Timeline

1. December 23, 2025: recruiter requested an updated transcript.
2. January 14, 2026: recruiter call covered the research area and interview structure.
3. January 26: completed back-to-back virtual behavioral and ML-domain interviews.
4. January 29: recruiter scheduled the onsite.
5. February 24: completed two back-to-back coding rounds at Google's MP3 office.
6. March 3: received the Team Match form and scheduled two hiring-manager calls.

The candidate chose the loop with one behavioral round, one ML-domain round, and two coding rounds rather than three coding rounds. The linked behavioral/ML report was inaccessible, so its unseen questions are not reconstructed here.

Onsite coding 1: simplified Mahjong hand

Given an array of 14 numbers representing a Mahjong-like hand, determine whether it forms a valid winning hand.

Practice: Validate a Simplified Mahjong Winning Hand

Onsite coding 2: dictionary prefix search

Given a dictionary and a prefix string, return every dictionary word that starts with that prefix.

Practice: Dictionary Words With a Prefix

The candidate described every interviewer as supportive and helpful. Their main advice was to communicate immediately when stuck rather than staying silent without an approach.