Projects
AI automation
Agentic LLM systems and evaluation, from a hands-off publishing pipeline to model selection at scale.
Experimentation and causal inference
Deciding what a change really caused, how long to run for, and which number to put in front of leadership.
Recommenders and personalization
Learning what users want and ranking what they see, at scale and under privacy constraints.
Modeling and evaluation
Predictive models and measurement systems built to be calibrated, explainable, and cheap to run.
Research and analysis
Studies that answer a question from data: regression, simulation, and quasi-experiments with a written conclusion.
Applied products and tools
Shipped products and utilities, built end to end and in use.
Combinatorial matrix theory and algebraic graph theory
Doctoral and postdoctoral research on matrices, graphs, and their spectra.
Computational neuroscience and signal processing
Analysis of brain and time-series signals, network dynamics, and spectral methods.
Teaching and outreach
Courses taught and materials built, plus interactive explainers and visualizations that make technical ideas approachable.