Meta London E5 Interview Experience - May 2025
Summary
I attended a Meta London E5 interview in May 2025, which included coding, behavioral, and system design rounds. I am currently waiting for the results.
Full Experience
I attended Meta(London) for E5 role from Berlin.
Screening: solved both
Unfortunately I rushed through the final interview 2 weeks after screening.
Codign 1:
- Vertical order traversal -> Solved it optimally
- Tic tac toe traversal -> Solved it optimally after initial non optimal solution
Codign 1:
- Group shifted string variabtion ( instead of shifting right, left used)
- Word Search -> instead of boolen, path needed. I had only 19 mins to solve this, so rushed with time optimal and non space optimal solution by keeping a copy of list. This problem is implementation heavy, so I couldnt dry run with test data but I wrote solution very fast after got the buy in. Not sure if this round backfire.
Behavioural:
- Interviewer seems nice, answered all questions, with followups. Felt I did very well.
System Design:
- Design uber ( interviewer is quiet throughout the interview )
- Drafted use case
- BOE -> messed up in some number for a particular use case ( fumbled for 10 seconds and moved on)
- NFRs -> perfect
- API -> Wrote all missed one, interviewer pointed out, covered it.
- HLD -> neat.
- Schema neat.
- At this point we had 15 mins left and I wrote possible deep dive areas.
- Dived deep on real time tracking once trip started(interviewer interested in this),
- Driver location update before order update
- Proximity search via redis geohash
Waiting for results.
Interview Questions (6)
I was asked to solve the Vertical Order Traversal problem. I solved it optimally.
I was asked a Tic Tac Toe traversal problem. I initially provided a non-optimal solution but then improved it to an optimal one.
I was given a variation of the Group Shifted Strings problem, where instead of shifting right, left shifts were used.
I was asked the Word Search problem, but instead of just returning a boolean, I needed to return the path. I had only 19 minutes, so I rushed with a time-optimal but non-space-optimal solution by keeping a copy of the list. This problem is implementation heavy, and I couldn't dry run with test data, but I wrote the solution very fast after getting buy-in.
The interviewer seemed nice. I answered all questions with follow-ups and felt I did very well.
I was asked to design Uber. During the interview, I drafted use cases, discussed BOE (where I fumbled in some numbers), NFRs, APIs (missed one which the interviewer pointed out), HLD, and schema. We then deep-dived into real-time tracking once a trip started, focusing on driver location updates before order updates and proximity search via Redis geohash.