Round 1: crypto order system, pretty similar to other write-ups already on the forum. I had to parse the input myself and write my own test cases. I got to level 2 — after I finished the tests, the output had no issues, but I wasn't given level 3.
Round 2: the "generate NFT" question, also already posted on the forum. Also had to define and parse the input myself — I used JSON. I finished the first two levels. For the third level I explained my approach and the interviewer seemed to agree with it, so I wrote it out the way I'd described, but when I finally ran it there was a bug I didn't have time to fix before time ran out.
Round 3: hiring manager round, standard behavioral questions, with a heavy focus on whether I had any AI/LLM related projects.
The feedback was that round 1 wasn't the optimal solution, and round 2 had issues with debugging/error handling. Thinking about it more, the data structure for round 1 was probably supposed to be Dict[State, Dict[str, Order]] to store orders. What I actually wrote was a separate dict for each state (live, paused, cancelled), with each state holding its own list of orders ([order1, order2, order3]). Honestly, that feels a bit like nitpicking to me.
Writing a workable solution plus tests was already a lot of pressure for someone just starting out with interviewing like me, and this company's bar felt kind of confusing. The first interviewer seemed pretty checked out — whenever I tried to interact and ask clarifying questions he seemed pretty unresponsive, and when I was thinking out loud it basically felt like I was just talking to myself... The Karat interviewers I've had at other companies were way more engaged than this one.
I interviewed for a mid-level position.
Discussion
Loading comments…