Markdown Files as Effective Memory for AI Agents
Why plain Markdown + Git is a powerful, transparent, and lightweight alternative to vector databases for agent memory — with real-world examples and practices.
Read MoreSenior Engineer at the intersection of AI, Data, and Product Systems.
I design resilient solutions that turn complexity into clarity, making technology work harder for people.
Why plain Markdown + Git is a powerful, transparent, and lightweight alternative to vector databases for agent memory — with real-world examples and practices.
Read MoreBreaking down monoliths into specialized agents for better reliability, maintainability, and scalability — with real-world examples from AI Research Weekly.
Read MoreMy work lives at the intersection of AI, data, and product engineering: think real-time ETL pipelines, agentic workflows that automate away headaches, and clean APIs that turn complexity into something people actually want to use. That’s meant cutting costs by hundreds of thousands, designing systems that replace hours of repetitive effort, and ensuring the kind of reliability that makes complex systems feel effortless.
What I enjoy most is turning chaos into clarity - whether that’s reverse-engineering legacy systems during an acquisition, streamlining processes, or making sure people can clearly see the story hidden in the data.
Outside of work - You can catch me out playing sports, exploring new places, or diving into a good book. I believe balance is important (all work and no play as they say) and inspiration often comes from the most unexpected places.
A Mastra-powered system that discovers new arXiv papers, enriches them with OpenAlex metadata, screens and ranks with agents/tools, and publishes polished weekly research briefs via a Next.js UI.
A secure authentication template with email-based allow list access control, built with modern web technologies and comprehensive security features.