the human behind the models 🦦

My favorite projects all start the same way: “I wonder if…”

Sometimes it’s a question about how a model learns. Sometimes it’s whether an LLM can solve a problem it probably shouldn’t. Occasionally, it’s because I thought, “That can’t be that hard.” It usually is.

I’m fascinated by the many ways machines can learn, reason, perceive, and make decisions—from machine learning, deep learning, and reinforcement learning to NLP, computer vision, robotics, LLMs, multimodal and agentic AI. I’m equally interested in what happens after a model works: why it made a decision, whether we can trust it, and whether we’re asking the right question in the first place. That’s what keeps pulling me toward causal inference, explainability, trustworthy AI, and high-performance ML.

If something catches my interest, chances are it’ll eventually end up here—a research project, a deep dive, a visualization, or an unnecessarily over-engineered feature that started as a five-minute idea.