AI Engineering

I got pulled toward AI engineering the way a lot of people did — watching, in real time, how much the industry was shifting and how much interesting technique was emerging around it. My professional entry point was leading the Microsoft Teams sentiment analysis product at PoliteMail: designing the pipeline that turned raw conversation data into insight was my first hands-on experience building with LLMs in production, and it opened up a space I wanted to keep exploring.

That exploration now happens mostly in my personal time, alongside my full-time role. I use Claude Code extensively — to learn faster, to organize my own workload, and to build tools that solve problems I actually run into. Two projects on this site come directly out of that:

DevOnboard solves a problem I’ve hit repeatedly in my career: a new hire or contractor joins a team with a complex stack, and even with good setup docs, getting them from zero to shipping their first safe change takes real time from senior engineers. DevOnboard scans a repo, runs an AI-driven gap interview, and generates a structured onboarding guide — aiming to cut that ramp-up time without requiring hand-holding.

Professor comes from a habit I noticed in myself: when I want to learn something new, I spend more time hunting down the right resources and structuring my own curriculum than I do actually learning. Professor acts like an AI instructor — it builds the curriculum, gathers resources, and assigns homework, so my time goes toward learning instead of research logistics.

Both projects are deliberately built outside my day-to-day stack — leaning into AWS and Python rather than the C# and TypeScript I use professionally. The goal isn’t just to ship something that works; it’s to show how I approach a problem from architecture through MVP, and to keep proving that out in areas beyond what my career has already covered.

Longer term, I want to work as an AI engineer — architecting and designing products that solve problems that genuinely weren’t solvable before AI, not just adding a chat window to an existing workflow. I’m also drawn to the idea of taking one of my own projects further, either growing it toward a real product or joining as a founding engineer somewhere. I’ll be honest that the founder path gives me real pause — I know myself well enough to know what it would cost my work-life balance — so it’s an ambition I hold with some hesitation rather than something I’m chasing blindly. What I’m sure about is the kind of work: I want to be the person architecting the solution, not just implementing someone else’s spec.