About Me
I'm a platform engineer with a background in computational neuroscience (PhD, WashU School of Medicine, 2020). I've spent my career at the intersection of complex systems and software—building tools that make intricate processes accessible, interpretable, and useful.
The question that drives my work: how does information processing architecture determine emergent behavior? I've pursued this question across neural circuits (EEG classification, closed-loop brain stimulation), ML systems (interpretability tools, ensemble methods), and simulated worlds (MUSE's biologically-grounded entity systems).
This lens shapes everything I build. Whether it's making neuroimaging pipelines visual and parallelized, turning ML from black boxes into interpretive environments, or designing simulation architectures where transparent rules produce complex behavior—I'm drawn to systems where you can trace cause to effect, where architecture matters, and where the "how" reveals the "why."
The tools I've built share a common thread: making complex systems accessible to experts with different backgrounds—and really, to anyone who wants to use them. MR.Flow replaces command-line neuroimaging workflows with a visual pipeline builder that auto-parallelizes execution to the user's hardware. Epoche turns ML from a black box into an interpretive environment—automated model comparison, ensemble inference, and tools that use what the machine learns to tell stories about the data. The barrier is usually intimidation, not capability.
Currently, I'm building MUSE—a simulation platform where characters emerge from transparent, scientifically-grounded cognitive models rather than scripted responses or black-box AI.
What I'm Interested In
My current intellectual obsessions, roughly in order of time spent thinking about them:
- Emergence in complex systems — The thread connecting neural circuits, ML models, and simulated ecosystems: how does architecture constrain and enable emergent behavior?
- Human-AI collaboration interfaces — Not replacement, augmentation. Designing systems where AI amplifies human judgment rather than substituting for it.
- Agentic AI and tool use — How do you build AI systems that plan, use tools, and collaborate with humans? What architectural patterns lead to reliable, predictable behavior vs. surprising failures?
- Interpretability and alignment — From SHAP/LIME to mechanistic interpretability to Constitutional AI. How do we make AI systems legible? How do we ensure they do what we want?
The core skill is experimental design. The questions matter more than the tools. I'll learn whatever tech stack the problem needs, but the fundamental capability is designing experiments that reveal how systems work. Read more about this throughline →
Why I Build What I Build
My background as a scientist influences how I build—from scoping features to iteration and testing. I track every decision along the way, not just changesets, but rich data that lets future me trace development back through time and analyze the codebase in new ways.
My scientific interests live at the intersection of embodied cognition, sensorimotor processing, and ecological psychology—how perception and action are inextricably intertwined, how context changes everything, and why the body isn't just a vehicle for the brain.
That domain expertise also shapes what I build. MUSE's heterogeneous GPU-native batch architecture—where rocks, trees, weather, and people are all processed together using a simulation loop inspired by closed-loop biological models—came directly from years spent thinking about how bodies and environments interact.
For more on how my neuroscience background shaped my engineering philosophy, see: Cognition Dx →
For a deeper look at how this question connects everything I've built—from neural circuits to ML models to simulated worlds—see Following One Question Across Ten Years.
Outside of Code
Music
Lead guitar in a band with heavy metal roots now bent toward alt-rock and jazz. However, we are prone to metal improv halftime shows (evidence below). My favorite sport is endurance tandem karaoke.
One of my favorite games! Guess which instrument I played in middle school band:
Gardening
Flower dork. Zinnias are my favorite. Somehow can't keep a succulent alive to save my life. The dahlias and tomatoes thrive. The cacti are never heard from again.
Cooking
Love experimenting in the kitchen. Allergic to following a recipe verbatim. Will turn a mild craving into a full blown tasting menu.
Snowboarding
One of my favorite places to be. Not shown: me not landing this and detaching muscle from my rib (with a very slapstick pop).
Worldbuilding
Dungeon Master crafting worlds for tabletop RPG campaigns. I am fascinated by the intersection of storytelling, the creative process, and cognitive science. Partially to blame for my obsession with databases.
Halloween
Go all in when the best holiday of the year rolls around. Will sew my own Princess Leia costume, but will outsource the buns. Zero shame.
Lupa
Black cat. Master keyboard obstructionist. Indiscriminate dice thief. Super fluffy.
Miscellaneous Mentions: lover of fonts, refuse to do things the slow way twice, camper, traveller, cycler, art installation enjoyer, and maximalist music of all genres(er) (modal walls of sound? sign me up!)
Currently
MUSE simulation platform and tooling ecosystem
Emory University (Lead Platform Engineer)
Children of Time by Adrian Tchaikovsky
Atlanta, GA (Oregon native, open to relocation)