The Interview
- it was an unusual interview, and very different than any communication or preparation material indicated it would be.
- it was supposed to be: 40 minutes, 10 for intro and building rapport, 15 minutes each with the software director and the engineering director, with the AI program director doing the meeting structure and observing, and one quality engineering person to help me. Oh, I was also the first person to go through this interview format, which they had never tried before.
This next part will be mostly meaningless to many but everything in preparation was: we're using React, .NET, Microsoft Azure, and starting to use AI tools -- we have Cursor but we don't use Claude Code. Goal is: I get to ask them questions, design solutions for how AI workflows and processes could help and provide opportunity. Outline the solution and propose a roadmap. In 15 minutes. with 3 minutes or so for the initial intro. Already feels improbable so I gave myself an edge, which was specifically allowed (more on that later).
What it actually was:
-
Started a few min late (now 36 minutes total time), and the initial 10 minute intro was being skipped in the meeting format. I managed to get a couple minutes of it but not many.
-
Software Director: Oversees multiple products, mix of legacy on-prem, hybrid, and modern cloud. Needs to continue modernization and support sometimes-frustrated customer base while still moving forward. Additionally, some tool and operating system incompatibilities are causing additional difficulties. I dig for more detail, needing to identify tech stack. I got what I could and gave her process improvement ideas that could apply broadly across the disjoint environment. She seemed interested in the answers. Friendly and easy to communicate with, she felt like someone I could definitely work with to create solutions. Natural conversation flow, but that doesn't lend itself to 15 minutes.
-
Engineering Director: Interesting guy, my age maybe a little older, really smart, more engineer type than corporate. He had two products, one of which he said was too complicated to even explain given the timeframe so we went with the other. Distributed system, cloud based, Ruby on Rails, not well separated, acquired from some startup. Totally different world -- not even close to anything else, even remotely. It would be like showing up to an interview for a Minister and having them expect a different religion. You'd find commonality and do what you could. It's pretty much like that. Also, this director was friendly, could have easily discussed engineering, and solved problems with and for him. But again, this isn't going to be a 15 minute start to finish chat.
-
Quality Engineer: My collaborator and quality engineer threw me for a little bit of a loop. She was extremely intelligent, and her groups were already using AI at a much higher level than I thought anyone there was doing, except for some research teams (which I was under impression were not connected). Using advanced tooling, defining agents and skills, probably at least looking at orchestration. Cool stuff that I really like as well, but a much different level than anything else led me to expect. In this case it just created one more unexpected dimension in a space where most of them were already unexpected.
It honestly sounded like there were some interesting (and tough) challenges, and people were using AI way beyond the initial "autocomplete and write some tests" levels and I would have had a good time helping them improve that process. The whole field is new and these tools evolve weekly if not daily.
I was disappointed. Felt like the format didn't really work, and considering it was their first time with this format -- it felt like a harsh evaluation. They also specifically praised about 8 different pieces of technical knowledge and social skills I displayed. I can't imagine who they'll actually pick if they stay with this format, but if that's really the way they want to do it, I don't think it will be anyone like me. It felt like a whole different culture from the first interview. The engineering director felt like the biggest exception to that.
About that 'edge' I mentioned, I figured what they were asking wasn't really possible in 15 minutes. So I had notes and outlines and such arranged on screens where I could see them without looking away from the people at the meeting. One one of those I had a project I built for Claude last night and this morning. I wrote blocks of about 20 different AI-driven process improvements, what they solved, what they provided, with specific examples. Also included estimates of how long they took 1st time, and later times (there is re-use so its' way faster after the 1st time).
Then I gave Claude a simple shorthand grammar where I could abbreviate the people (SD software director, etc), issue types, and a couple other things -- it would pick the right pieces and spit me back a roadmap with specific items split into phases, with estimates and any special benefits like cost savings, which it would calculate custom if I included team size in the shorthand. and it worked! It was essentially a "natural language state machine that accepts a custom shorthand".
Through this whole process I also knew that if it didn't come to be, it wasn't supposed to. I knew this because I did all of the things I was supposed to -- I wrote it down, tarried toward the goal and did a possibly-extreme level of preparation. At some point, I trusted in God that if it were meant to be it would be. There are a couple things I could have done or said differently but honestly it was a hell of an attempt. Given all that, if it didn't happen it wasn't the right thing, but I still wanted the resolution of having it be.
But here's what became clear after I thought about it
My only "this might not work out" thought going in was — what if it's too corporate and I just don't fit? I was willing to fit that part of the mould if needed to get an AI position. But I don't think it's where I belong.
Intense engineering I can do. I'm not against sleeping on the lab floor, and I find appeal in the fact that there are places where that still happens. I miss the weirdness of labs in the middle of the night, the crazy geeky culture of doing stuff not just because it makes money, but because it's honestly just so exciting to see where the possibilities can go.
I want to be somewhere that the work itself is the point. Where the people around me are there because they can't imagine being anywhere else. Where the hard problems are actually hard — not because the process is broken, but because the frontier is real.