Red flags from the beginning, but I didn't heed them. Initially, I was reached out to by a recruiter (the recruiter was a very wonderful person and great to work with – the only good part about my Capital One interview experience). I decided to give the coding assessment a shot. I passed that, and instead of being able to talk directly to a hiring manager to learn more about the specific role, they moved directly onto a "Power Day" interview, a ~4.5-hour time commitment.

It turns out that Capital One doesn't hire for specific roles. They bring you in and then have you interview with four different people who then, if you pass the interviews, assign you to a group to work with. While I can understand that Capital One is a large company and has many openings and groups, this is a terrible format for hiring individuals who have niche skills or interests. Unless someone is generally thinking, "Oh, I want to work for Capital One doing anything," then this process wastes time on both sides.

I have a set of skills and an interest in how to grow and apply them. From my reading, I thought Capital One would have spots available for me to pursue that growth. The hiring managers weren't specific, and the interview questions focused on generic terminology versus discussing applied experience.

I forgot a couple of my technical terms (yes, that one is on me for not preparing better for them) that I hadn't used or discussed in a few years. Then, I was not able to critically discuss my applied knowledge and projects to a deep extent. I had a hunch I wouldn't make it past the Power Day (I didn't; the technical interview folks said that they had to ask too many guiding questions), and I was intending to decline the next round if that was in the cards.

The format was:

• 2 Technical interviews with hiring managers who do some ML to a certain extent
• 1 Behavioral interview
• 1 Case Study

As stated above, the technical interviews didn't allow me to discuss my work experience. Interviewers were nit-picky about exact technical terms for some broad concepts. One allowed me to discuss a project I had done, but that happened toward the end of the call. The second (in a more customer/production group) presented an ML-type problem and asked me how I would go about tackling it. He got really in the weeds about some of the computation optimization, which isn't something I could answer offhand.

The behavioral interview was good and straightforward. The case study didn't relate to ML at all and was poorly structured, but the interviewer and I had a good conversation.

TL;DR: If you want to work at Capital One for the sake of working there, by all means, go for it. For the most part, everyone I spoke with was very nice and excited about their jobs. If you are interested in a specific role, they don't exist. You'll have to go through around 7-8 hours of your time talking with a recruiter, taking a coding exam, then taking part in a broad Power Day interview before you get specific details about a role. It was a waste of my time, or maybe it saved time so that I'm aware of this in the future for other employers.

Here are a couple of interview questions I can remember:

• What is the bias-variance tradeoff?
• What's a GBM and an example of one?
• What is underfitting and overfitting?

In a case study interview, you might be asked to discuss how you would optimize the cost of cloud compute machines.