Every blog needs a first post. This is mine.

I’ve spent years doing research and building ML systems — writing papers, filing reports, shipping models to production. But the kind of writing I’ve done the least of is the kind where I just think out loud.

That’s what this blog is for.

What I’ll write about

My work sits at the intersection of a few areas I find genuinely fascinating:

• Large language models — how they work, how to build systems around them, and where they’re headed
• Agentic AI — multi-agent pipelines, context engineering, the gap between demos and production
• Bayesian reasoning — probabilistic thinking applied to real problems
• Structural health monitoring — my original research area, which I still think is underexplored by the ML community

I’ll write about things I’m building, papers I find interesting, and ideas I’m still working through.

Why now

I recently built agent-as-judge — a multi-agent pipeline that evaluates documents against industry benchmarks using CrewAI and a local Llama model. It was a fun project, and I realized the most interesting part wasn’t the code — it was the design decisions along the way.

I want a place to document those decisions.

What to expect

Short posts, technical depth, no filler. If I don’t have something worth saying, I won’t post.

Let’s see where this goes.