YOE : 5
Current job : ML Engineer
LC Solved : ~300. Easy 50 / Med 180 / Hard 70

Decided to take onsite in two separate days.
Day 1 : Coding x 2, ML Coding x 1, Behavioral x 1
Day 2 : ML System Design x 1 (60min)

Due to NDA, I won't share actual interview questions.

Phone interview : LC Medium question with minor twist. Solved in 20 mins, including writing unit tests, comments and documentation. Brought into onsite.

Onsite interview - Day 1
Coding session 1 : LC Medium-Hard question with minor twist. Due to initial miscommunication, I solved wrong problem until ~10 mins. The interviewer led me to the right path. I was able to reach O(1) solution and write unit tests with edge cases and random test cases.

Coding session 2 : LC Super Hard question. Similar to LC question but never seen exactly the same question in LC. Barely cobbled writing a full code with suboptimal solution. The interviewer asked space/time complexity and I answered correctly. The interviewer said, "Well, this is not completely correct as you know, but you're the first one who reached this far when I gave this problem several times..."

Machine Learning Coding Session : Coding question requiring strong machine learning background. I'd say its difficulty was Medium-Hard in LeetCodish standard. The interviewer asked to parallelize my code in multithreaded version. I've managed to do it. The interviewer then asked bunch of machine learning knowledge questions in rapid-fire salvo. I was able to answer 70%+ of them correctly, but was not able to answer the rest.

Behavioral Session : Just standard behavioral questions. The interviewer looked satisfied.

Finished Day 1, the recruiter told me that remaining Machine Learning System Design interview would be the most important factor to decide the level of offer (L4 or L5 or Reject). Prepared hard.

Onsite interview - Day 2
Machine Learning System Design - A broad, medium-hard problem. Definitely nailed this one.

Waiting for result.

EDIT: The recruiter told me that I've been rejected for two reasons:

1. In "Machine Learning Coding Session", I was not able to correctly answer ML knowledge questions everything.
2. There was only one headcount in my region, and the hiring manager decided someone else to hire.

This feels like shit but I have to face the truth. I have to study ML knowledge in-depth.