[Jemaclus]

The CSPA thing is interesting to me because the founder/CEO was formerly the head of engineering at Crunchyroll, and interviewing at Crunchyroll was one of the worst interviewing experiences I've ever had.

Let's just say I'm skeptical. I hope it works out, but...

MacD looks interesting... but I'm kinda surprised YC would invest. A Mac n Cheese company? Very strange.

I'm very excited about the biotech space, so looking forward to seeing how that shakes out.

[cspa]

Sorry to hear it was that bad! :( I'd love to hear your experience, especially if I was directly involved -- you can reply or email me anonymously at [e-mail]. If you know anything about Crunchyroll's early culture, I always made a point to listen to feedback and improve ourselves. It'll be the same at CSPA.

Crunchyroll's interview process definitely was not consistent, which is why I spent my final two years there trying to improve our processes. I learned a lot from it, and that's why we're doing CSPA :)

- James Lin ([e-mail])

[Jemaclus]

I have no idea if you were directly involved, unfortunately, which was part of the problem. I described my experience in another comment: https://news.ycombinator.com/item?id=17795308

[krackers]

>Interviewing at Crunchyroll was one of the worst interviewing experiences I've ever had

Can you elaborate?

[Jemaclus]

I was a Senior Software Engineer at the time.

I thought the coding challenge was fun. They basically give you a URL, and the URL contains a list of numbers, which you append to the URL. Visiting each URL+number returns either a list of other numbers (which you keep traversing), a FAIL message (dead-end, of which there are many), or a SUCCESS message (only one). You task is to traverse the entire tree, determine how many dead-ends there are, which number is the successful end, and which path is the shortest path from your starting node to the success node. It was pretty fun, and the solution wound up being basically a breadth-first search or a depth-first search (pros and cons for each).

(You can do the challenge here, if you want: http://www.crunchyroll.com/tech-challenge/roaming-math/yourn...)

I submitted my answer and got set up with a phone call. I don't remember anyone's names, but this guy was touted as a PhD and a genius, so I was excited to talk to him. Over the phone, he gave me another challenge: "Imagine you're Photoshop, and someone selects a pixel with the Fill tool. How do you know which pixels to fill?" As you can imagine, this turned out to be another breadth-/depth-first search problem. You look at the surrounding pixels, decide whether they're the same color or not, and then proceed to the next level and repeat.

I was offered an on-site, which I accepted. I arrived and sat in the lobby. They have a giant big-screen TV that was showing anime. Makes sense, since CrunchyRoll is an anime streaming service. The current show had a woman eating in a restaurant with some men. I couldn't understand the words, but it was clear that she was having orgasms while eating, and the men were watching. At the time it was mildly amusing, but in retrospect, it was quite inappropriate.

Eventually, I was shown to a room with a whiteboard. The first person (an engineer) comes in, says "Hi, I'm Jack. So you have M trains and N stations..." and starts writing on the board. No idea who this guy is or why he's talking to me. I, being nervous, tried to follow along. His question was ultimately another search problem. Due to poor time management, I didn't have time for questions. He nodded and walked out to get the next person.

Second engineer comes in. "Hi, I'm Brad. You have M dogs and N cats..." and jumped right into his problem. No context, no anything. Yet another search problem. At the end, there was no time for questions.

Third guy comes in, same thing.

Fourth guy comes in, he actually introduces himself. He's the head of Product (or at least some Product Manager). He asks me a couple of questions, then gives me a coding challenge on the whiteboard. Another search problem.

Fifth guy comes in, the PhD guy comes in, says "Hi, I'm Kevin. You have..." and launches into yet another search problem (again!). Having done six of them by this point, I breezed through it and finally had time for one question. I said, "You guys must do search problems a lot here at CrunchyRoll."

He cocks his head to the side, looks at me like I'm crazy, and says, "No, never."

The sixth guy comes in and gives me something different! "You have a box and a bunch of random objects. What's the most efficient way to pack it?" which, of course, is the Knapsack Problem, which I believe has not been fully solved. But I do my best.

At this point, I'm over it.

I know CrunchyRoll does not fill boxes. I know they don't do any of the things they're quizzing me on. I don't know anything about the business other than what I've read online. I don't know what any of these people do. I haven't been given any context about anything, and when I did manage to ask some questions, I got nonsensical answers. I don't know which team I'd be working on or what role I'd take over. I don't know which of these alleged geniuses would be my coworkers (or god forbid, my boss). I don't know what problems CrunchyRoll's engineering team is trying to solve. I don't know their tech stack, their culture, or their company strategy.

I was not offered a position, but even if I had been, I would have turned it down. It was pretty bad across the board.

Ironically, the next company I worked for did some work on behalf of CrunchyRoll in the data engineering space. Thought that was funny.

I had a similar experience at Zenefits the next day, but halfway through the interview I declined to proceed further once I realized it was the same thing.

On the positive side, I've interviewed with a dozen other companies in the intervening years, and most of them have been fairly positive.

[krackers]

>The current show had a woman eating in a restaurant with some men... At the time it was mildly amusing, but in retrospect, it was quite inappropriate.

Probably was Shokugeki no Soma

The first few rounds don't really seem too dissimilar from other firms (in that they'd ask a bunch of random algorithmic problems that you will never likely come across in your job). That being said, six consecutive rounds of interview with these sorts of problems does indeed seem a bit too much.

[Jemaclus]

It wasn't necessarily six consecutive rounds with "these sorts of problems", it was six consecutive rounds of _the same problem_ just worded differently. And the actual questions were less the terrible part as the fact that interviews work both ways. An interview is a chance for _me_ to learn about _you_ just as much as it is for _you_ to learn about _me_. A very good software engineer in a city with very high demand for software engineers has a lot of possibilities. CrunchyRoll is not the only person asking engineers to spend their entire day in the office. You can't treat an interview as an obstacle course that the candidate has to overcome.

Crunchyroll showed zero interest in learning anything about me other than whether or not I can search a tree. They showed zero interest in letting me learn anything about the company other than I don't want to work there. They showed zero ability to collaborate and coordinate prior to the interview. They showed no interest in me personally as a potential team member. They showed zero ability to recognize that I gave up an entire day (and likely lied to my boss about why I was missing work that day) to come into their office and... waste my time? Yeah, not great.

Conversely, I've been involved with dozens of interviews at other companies that ask algorithmic questions that were pleasant, interesting, and an actual two-way street. There were plenty of companies whose interviewers seemed to actually know my name, and who spent time trying to convince me that I _should_ work there.

IMO, if you bring candidates on-site and you're only asking them algorithmic questions, your recruitment funnel has failed. You shouldn't bring anyone into your office unless you're pretty sure you're going to hire them. That means determining whether or not they're competent _before_ they come in. The on-site should be a validation of what you already know (e.g., verify they actually did the coding challenge you sent them and that they're the same person from the phone call(s)), a confirmation that their temperament and personality are compatible, and then selling them on joining the company.

To be clear, I don't have a problem with algorithmic questions, but more than one or two is a waste of time.

And yes, this pattern is sadly common in SF, but common does not make it okay.

[cspa]

It sounds like this happened after I left Crunchyroll in 2015. The process I had in place was definitely not like that :/

Well FWIW, I personally wrote the coding exercise that you enjoyed (http://www.crunchyroll.com/tech-challenge/roaming-math/yourn...). I thought it was a good mix of basic algorithms (tree traversal) as well as some practical knowledge (CURL/http). I'm surprised they are still using the same one :)

[Jemaclus]

My interview was in June 2015. Maybe you had left by then. It did seem disorganized.

I did really enjoy that coding challenge, though. Good job with it. It is the one thing I did like about the whole experience, and I still share it with people when they ask about what a good coding challenge might be. Tree traversal isn't particularly useful in web dev world, but it's a pretty intuitive task, so I think it's gives good insight into whether a programmer can reason logically about a task.

I'd rather see real-world challenges, though. Sometimes that's hard, though, depending on the company.