I contacted them on Stack Overflow and was asked to solve an online challenge. Once you start, you have 72 hours to complete it. The task required parsing exchange orders and executing them, or something like that (basically building a mini stock exchange; I don't recall the exact details).

I really liked the task; it was interesting, and I invested a considerable amount of time. The result of all my time devotion was completely ridiculous: a three-line email with this reply: "The results of your testing did not meet our minimum requirements to move forward in our process." That's all.

I had worked before that in three to four companies (each was over five years) and in all of them, I was one of the top-performing software engineers. So, I was quite confident about my abilities. After investing three days of work, it didn't feel right that the company asked people to spend three days of their work time for free and then didn't even bother to properly provide feedback on what they found wasn't okay with the work submitted.

So, I emailed them and asked that I had invested two to three days of work and that extra feedback would really be appreciated. So... I don't know what kind of people work there (we all know that every company believes that they have the best people, but the truth is that they don't), but the feedback from their "engineers" that was the basis of the rejection (copy-pasted from their reply) was:

• Very inefficient parsing; reads data one byte at a time.
• Misuse of shared_ptr; there was nothing shared in the code.
• "typedef" instead of "using".

It doesn't say that the task didn't produce correct results, or that something was clearly wrong. The only point for a large task that might have non-cosmetic weight was "Very inefficient parsing; reads data one byte at a time," and this was a completely wrong judgment. They didn't even bother to understand that it was, in fact, very efficient and did not read data one byte at a time as they naively thought.

Experienced software engineers know how to write proper code, and very often, something that may look "inefficient" is, in fact, translated by the compiler to the most efficient form. That is, some engineers naively think that they can do better than an optimizing compiler (it wasn't about SSE or anything like that in the task, just plain structs loaded from memory).

When I explained in the reply that their evaluation was not correct and why, I got no reply.

That was the worst interview ever, or rather, the worst company that I applied to. AKUNA: you should not treat potential candidates with such disrespect where you don't even bother to invest 15-20 minutes of your time, while you ask people to spend three days on your tasks. I interviewed in many places and got offers pretty much from all the top tech companies from Silicon Valley and from some investment banks for their algorithmic trading engineering jobs (MS). All of them, at the final interview stage, take four to five hours of your time, but each of these hours, they put their interviewer in front of you. That is, they also invest the same amount of time from multiple engineers.

The offer that I eventually accepted was from a high-growth, very successful startup (last year, it was listed among others that reached unicorn status, over $1B valuation). As usual, I'm also one of the top-performing software engineers in that company. Even though that was industry-unrelated to my previous experience (just like I had no idea what algorithmic trading was when I interviewed at Akuna), in that company, my salary almost doubled, and I was promoted after 2.5 years to technical director level (the company itself also grew from 100 to 500 people).

I know most CEOs of companies do read all their interview feedback, and this message is directed to them: could it be that your B-level players only want to hire C-level ones?

write trade execution engine