About
The Harness is where I write about harness engineering, agentic systems, and the work of making AI useful in real engineering practice.
Hello, I am Theodoros Galanos. I have been working at the intersection of AI, design, and engineering for the better part of the last 12 years.
People who know me know that I tend to be opinionated about where the future is going, and equally interested in how that view changes over time. This site is where I try to make those ideas more concrete: a place to write about agentic systems, harness engineering, and the harder middle layer between model capability and reliable real-world work.
My interests sit around the design of systems that need to do more than produce plausible answers. I care about environments, tools, verifiers, workflows, evaluation, and the control surfaces that let people direct useful autonomous behaviour without losing accountability.
The Harness exists to advance that work in public. Part of that means research: trying to make vague claims about agents more concrete through better tasks, better benchmarks, and better system design. Part of it means building: turning ideas about orchestration and reliability into actual tools, workflows, and experiments. Part of it means education: writing things clearly enough that other builders and domain experts can use them. And part of it means community: helping create a sharper conversation around what harness engineering is, why it matters, and how to do it well.
If you want to follow along, you can subscribe by email from the footer form or use the RSS feed. If you want a broader snapshot of my work, the external links in the header point to my LinkedIn and X. Otherwise, the archive stays simple on purpose.