I interviewed for Staff Software Engineer role. Mountain View, CA

• Phone Screen (75 mins)
  • 1-2 Behavioral questions
  • Standard coding problem. Would be leetcode medium, don't remember the exact prompt, but it was straightforward. Some followups on scalability of the system, issues we might get etc etc.
  • 20-25 mins of AI questions. Since my background is in building AI applications, lots of questions on these. Another discussion on AI native vs non AI native apps. What are LLMs, how do you use AI in your day-to-day life, what are some dangers of AI etc.
• Onsite
  • Craft Demo (75 mins)
    • A 3-5 slide presentation of your background, hobbies, proudest project. (10 mins)
    • They give you a repository to setup in advance. You can play with it beforehand. It is a simple SpringBoot application with some endpoints defined to fetch some data from a datastore. They give you some user stories to work with (add more apis to support these). Focus is more on how you solve problems, what all you will test, how will you design classes etc. They don't care too much about running and verifying or adding unit tests, if you do this that will definitely give you more points. I completed the basic implementation, tested it with Postman but could not write tests. They said its ok to use chatgpt to do this, but focus was on what parts to test and how to test in unit test not the syntax.
  • System Design (60 mins)
    • This interview focused more on the scalabitlity and high level design of the problem solved before. It was an in-depth discussion on schema design, pros and cons of each proposal. There were a lot of follow ups on expanding the original service to add more features, how the db would change, what would be the query performance, how can certain databases be used for optimizing searches, how to add indexes etc. This was an in-depth conversation.
  • AI Accessor interview (30 mins)
    • This was adding an AI related follow up to the service built in round 1. The service was also using an LLM api and I had to write an API to support the AI use case related to recommendations. This was a discussion around RAG, LLM hallucination and prompt engineering.
  • Hiring Manager (30 mins)
    • This was a standard hiring manager interview with the VP. Behavioral questions, what excites you at work, what are your superpowers etc etc.

Recruiter reached out after 2 days to say that the team would like to give an offer.

Looking back, I think these interviews were the best, focusing on the right things e.g. More design / architecture focus for Senior/Staff candidates instead of too much focus on coding. Honestly, I did not complete the coding (e2e with unit tests n all) for both the rounds, still the feedback was positive. The interviewers were a delight to talk to.

Preparation Post: https://leetcode.com/discuss/post/7368458/coupang-intuit-salesforce-oracle-c3ai-se-n0gq/