About

I'm an engineering manager and machine learning engineer with a focus on NLP and ML systems design, building high-trust teams, and strong software engineering practices.

In practical terms: I build systems that do things with ML that ML is actually good at, and I also build and tend the human systems that make it go.

I lead the engineering team at Hidden Door, where we're building a platform for fanfic-meets-roleplay interactive story games.

Evergreen interests (I'll talk about this any time!)

Using the right ML tools in the right places
Finding the ayyy lmao in startup engineering
Clear, effective communication/collaboration
Management, especially remote and multidisciplinary
Testing & type-checking in Python

Newer interests (What's on my mind lately?)

Using AI code assistants in ways that don't suck (for the code or the human)
Where culture and structure meet re: the above, and how to adjust those for better outcomes
Observability

Some things I've done

I've used classical machine learning for NLP tasks and classification/prediction, implemented time- and bug-saving infrastructure in large and poorly-tested codebases (think testing, observability, type-checking, and the like), and spun up your typical AI engineering prompting/ testing/ evaluation/ iteration setups. I have considerable applied experience with LLMs, both in the products I build and as a tool in my own work.

I'm known for clearly communicating this work and collaborating with teammates who have a wide range of backgrounds and scopes of concern.

I've built major parts of the platform and ML systems for narrative games at Hidden Door, balancing a complex set of constraints and requirements—both human and technical—even before you account for the fact that creating a good interactive story is incredibly difficult.

I've also fixed or helped fix some very tricky sociotechnical challenges.

Before my engineering career, I did a PhD in evolutionary biology at Harvard, where I studied why horses evolved one toe.