The interview process began with a phone screen, where my experience and background were discussed. The interviewer inquired about my motivations and passions.

Next was a technical interview. The first technical interview went quite well. The interviewer was very friendly and approachable. The question focused on the filesystem. They adhere to a specific script with checkpoints and ratings. It's crucial to communicate your thought process and assumptions clearly. The primary focus appeared to be on assessing algorithmic knowledge, specifically BFS/DFS.

I received a call from the recruiting team, who informed me I was precisely what they were looking for. They stated I performed significantly better than most candidates and expressed their eagerness to extend an offer.

Following that was a second technical interview, which involved another task: the ID allocator question. This interviewer was also very pleasant. By this point, I understood they followed a script. Both interviewers mentioned something to the effect of, "It may look like I'm not with you, but I'm just writing down your answers." I passed this interview as well. This stage has three levels; reaching the last one involves considerations of space constraints (bits/heap).

Once more, the recruiting team contacted me. They confirmed I had passed this stage and were very pleased with my performance. They explained there was one more interview (or possibly two, with HR being the last) before a final decision.

Subsequently, a third interview was scheduled. Unfortunately, the interviewer experienced technical difficulties and could not connect. Someone else was brought in at the last minute. This interviewer presented the web-crawler question. The objective was to demonstrate proficiency with threads. Regrettably, this is where I faltered, as I did not handle the task effectively. The interviewer repeatedly asked if I knew the algorithm I was using, but I misunderstood the question, believing it referred to thread ordering. Despite showcasing my BFS knowledge in the first interview, progress from previous stages seems to be disregarded, requiring you to demonstrate your understanding anew in each interview.

After some time (approximately a week), I received a call informing me I had not passed. The feedback was a "very weak no," and it was mentioned that the last interviewer was particularly stringent with all candidates. However, they still considered my performance to be quite good, though not quite good enough for Dropbox.

Overall, I would describe the interview process as very straightforward. The questions are relatively easy, but effective communication of your thought process is paramount. The interview questions are intentionally ambiguous, so articulating your assumptions and the rationale behind them is essential.

Q1: Given a root directory, find all duplicate files and return them. Extension 1: Assuming you used hashing, can we minimize the number of hashing operations?

Q2: Write an ID allocator that can allocate and release IDs. Extension 1: Can we make it more space efficient? Extension 2: Can we make it faster?

Q3: Given a function that fetches the HTML of a webpage and a function that fetches links from the HTML text, write a web crawler that will find all the links. Extension 1: Can we make it threaded?