I applied online. The process took 4 weeks.

I interviewed for a senior ML/AI engineering role. The loop covered coding, an ML-specific round, behavioral, and ML system design. The process is well organized, and the recruiter was a huge help. I've never seen that level of involvement. The interviewers are real professionals and sharp people.

Coding is a bit old-fashioned: medium/hard LeetCode-style problems. Most of the time the round is an open-ended task with no concrete input/output. For example, you discuss how you'd design a data structure, implement a baseline, and then discuss optimisations. One of the pros for me was that in system design you pick your own domain of specialization as the main topic. Overall, it was a great interview experience.

The process covered several areas:

• Coding: graph structures, arrays of intervals, and prefix sums.
• ML: a discussion of LLM inference, plus a bit of coding around it.
• Behavioral: stories from past experience, mapped to Googleyness values.
• System design: design a distributed system built around LLMs.