Google "Level targeting Interview", how to prepare?
YOE : 6
I applied to Google last year and was rejected at the onsite. The target level was L4 or L5, ML software engineer. The interviews consisted of 5 sessions, 2 coding, 1 ML design, 1 ML domain knowledge, 1 behavioral
The recruiter said : "You did well in others but ML domain knowledge performance has some doubts. Unfortunately the headcount was one and the hiring manager decided someone else to hire. You can apply after 8 months"
Recently they contacted me, saying "The 8 months have passed, will you take interviews". Sure, why not. Since last onsite performance was not terrible, skipped phone screening.
Me : "What is the target level?"
Recruiter : "L4/L5".
M : "Can I take the interviews as L5 target?"
R : "Possible, but then you have to take the 'level targeting interview' in advance".
M : "??? What's that? I've never heard of that before."
R : "It's new one for cases like you. You'll take one coding interview, and based on the result of that, the target level of the remaining interviews will be set. If the interviewer feels 'your seniority is enough', the target level will be set as L5, otherwise L4."
M : "What's the difference between this and phone screen?"
R : "Phone screen result doesn't count in packages that will be sent to HC. And this is counted as onsite, so after this, you will take 4 remaining onsites, not five. Once the target is determined to be L5, if you can't make the L5 bar, there is no L4 low balling, you're just out in that case"
M : "Got it. So how can I 'show my seniority' in coding interview? Aren't this kind of things usually decided by design interview?"
R : "You just have to perform very well."
I have no clue about this, could anyone help me how to prepare this kind of coding interview?
During preparing I'm trying to write clean, idiomatic, robust code beyond LeetCodish often unreadable codes. Am I on the right path?
#swe #google
comments
1) your way of thinking - how you arrive at the "optimal" solution.
2) once you discuss the pros and cons of the brute force vs the optimal how do you validate your solution?
3) implementation: apply OOD concepts (separation of concerns, inheritance, polymorphism, etc..)