I build AI systems end to end, from Bayesian models to LLM agents. Here is some of that work.

Choosing a Model for Every Coding-Agent Task

A small router can help coding agents choose a model for each subtask. The useful engineering is in the guardrails around that choice.

Choosing Priors in Bayesian Analysis: A Gentle Guide

A practical, example-driven walkthrough on how to choose priors in Bayesian modeling with real data, PyMC code, and a few lessons learned along the way.

Building a Virtual Theremin with MediaPipe and Pure Data

Using hand tracking to create a virtual theremin with MediaPipe and Pure Data.

Bayesian Filtering for Audio Denoising

Applying a Bayesian Filter using RxInfer for denoising audio data.

Kalman Filter and RxInfer for Structural Dynamics

Implementing an AKF in Julia and RxInfer for Structural Dynamics analysis.

Bayesian Classification and Survival Analysis with NumPyro

Applying Bayesian Classification and Survival Analysis to the Cirrhosis Patient Dataset.