Blogs
What I Learned Building Multi-Agent ML Systems at a Startup With No Established Infrastructure
Most ML engineering advice assumes you have clean data pipelines, dedicated MLOps teams, and well-scoped problems. I had none of that. The lessons that stuck have almost nothing to do with model architecture.
Lessons From Running a Real Business That Made Me a Better Engineer
I've hosted regional fencing tournaments for three years through CozmX Fencing. Real money, real logistics, real failure modes that no rollback command can fix. And I'm convinced this made me a sharper ML engineer than any Kaggle competition ever could.
Things I Believe That Most People in AI Would Disagree With
Most ML engineers would be more effective if they stopped reading papers for six months. Agents are being dramatically over-architected. The best preparation for AI engineering isn't computer science. Each claim should make you uncomfortable — then convince you I might be right.
Position-Dependent Vulnerability: How Early Misinformation Derails LLM Reasoning
This research explores how the position of misinformation in prompts affects large language model reasoning capabilities. We investigate whether LLMs are more vulnerable to misinformation presented early in a prompt versus later, and the implications for prompt engineering and AI safety.
More blog posts coming soon...