I applied online. They sent me a data challenge that they said I should expect to spend 2 hours on, which is unrealistic because the data was complex and you have to write up your solutions.

The instructions were very vague, and I took the shorter interpretation of those instructions. After spending approximately 10 hours on the assignment and write-up, I sent in the assignment.

A week later, I got a form letter saying: "We appreciate your interest in Flatiron Health. Unfortunately, there's not a strong fit for your background at this time."

That was kind of rude, considering I'd been communicating with HR.

I emailed HR and asked why they sent me a form letter and for feedback on my code. They apologized for the form letter. HR said that my solution didn't meet their bar in terms of answer correctness and code quality.

So I asked for more feedback because I wanted to understand my mistakes, as I thought I was very thorough on the challenge.

Eventually, someone else from HR got on the phone, and they offered me very generic feedback that didn't really seem to pertain to my solution. They talked about their "holistic approach" to evaluations and said something about how I didn't do something that I did. In fact, I offered several different explanations of this topic because the instructions were so vague.

But it seems like ultimately I got the answer wrong because I took the wrong interpretation of half of the vague questions. So, it sounds like they were evaluating not my answers but how my answers matched up to their unstated interpretation of the vague instructions.

Also, my code was "unnecessarily complex." But I don't see how someone is supposed to write beautiful, clean code and take the more time-consuming interpretation of their questions in an unpaid "2-hour" coding challenge.

I don't think Flatiron Health's HR practices are particularly well thought out. They seem confused and disorganized.

This data challenge looked at three different cancer drugs on breast cancer and colon cancer patients, separated into 13 different sub-types of cancer. They asked about duration of treatment and things like that.

It was challenging to interpret the vague questions. It seems like when they asked you to evaluate drug duration by cancer type, they didn't mean breast and colon cancer but rather these sub-types of cancer, but that wasn't stated in the instructions.