Google 2026 PhD New Grad AI/ML SWE: Tech Rounds and Bay Area Onsite

Role: Software Engineer, PhD New Grad — AI/ML
Result: Entered Team Match and later matched
Report time: August 2026
Application: Applied online

Timeline

1. Applied online.
2. Received a phone screen about 60 days later.
3. Completed the ML and behavioral technical interviews around day 90.
4. Received a positive result and scheduled the onsite around day 105.
5. Completed the onsite around day 120.
6. Entered Team Match around day 135 and later reported a successful match.

ML and behavioral interviews

The ML round focused on fundamentals tied closely to the interviewer's own background. The candidate did not know the interviewer's identity in advance. When an answer was unclear, the interviewer offered hints and sometimes explained the correct answer. The source did not preserve the exact ML questions.

The behavioral round used basic questions and did not focus on collaboration, failure, or conflict stories.

Onsite coding 1

Given an array, choose two indices that maximize the sum of all elements from the first index through the second minus the sum of the elements strictly between them.

After clarification and examples, the candidate recognized that the expression reduces to the two endpoint values. They discussed brute force and prefix sums before presenting the simpler mathematical interpretation. In the comments, the candidate confirmed that scanning with largest and second_largest variables was valid.

Practice: Maximize Two Endpoint Values

Onsite coding 2

Each employee record contained an ID, name, on-call start time, and on-call end time. Return each time segment together with the number and names of the employees active during that segment.

The candidate quickly recognized the interval/heap pattern. The interviewer then asked for exact time and space analysis for every operation and data structure, down to expressions such as O(n + k + h), with every symbol defined in code comments.

Practice: Build a Combined People Schedule

Onsite experience and follow-ups

The two coding rounds were back-to-back with no break. The candidate emphasized that Google evaluated communication closely: clarifying assumptions, walking through examples, explaining while coding, and defending complexity choices.

Google offered Bay Area or Seattle for the in-person onsite, with travel, hotel, transportation, and meals reimbursed when itemized receipts were kept. A later comment said New York could also be available. The candidate did not know any interviewer's identity before arriving.

The lunch conversation was non-evaluative. The candidate later updated the thread to say that they had successfully matched with a team.