I went to the onsite around October or November last year — sharing my experience getting rejected.
I'm a US undergrad. I went through one coding OA and two math online interviews before getting to the onsite. The math rounds covered some probability and combinatorics questions — you need a solid grasp of the Central Limit Theorem and basic probability. The onsite has four rounds: programming, data analysis, system design, and math. I was assigned to go to the office in the morning — two rounds in the morning, two in the afternoon — with lunch at the company in between. If your results in the first two rounds aren't good, they might just cut the interview short right there.
Different people get different round orders — I drew programming and data analysis for the morning. Each round was an hour.
The programming round: they gave me a C++ codebase and asked me to add a feature. It was a trading system with a client and a server that already had add and cancel — I needed to implement modify. Basically you had to find your way around the codebase, get a rough sense of the overall framework, then copy-paste and tweak things until it worked.
The data analysis question was drawn from a question bank — the difficulty all seemed about the same. I got a heart disease prediction problem, with a bunch of features like age, weight, height, and so on. You basically needed to be comfortable with pandas and seaborn for plotting, and you were allowed to check the documentation. I'd prepared beforehand, but since I'm an AI major I wasn't that comfortable with pandas, and I wasted a lot of time on basic code. At the end the interviewer asked if I normally don't use Jupyter Notebook, and that's when I felt like I'd failed.
They told me after lunch that I didn't pass. Hoping this helps more people.
Discussion
Loading comments…