Blog

Why autonomous AI agents fail in production. The problem isn't bugs, it's the mismatch between non-deterministic LLMs and business processes that need predictable results.

How to scale the deterministic core principle: build a team of specialized agents instead of one monolithic AI. Each agent has a controlled role with specific tools and context.

Analyzing the recent Tea App data breach that exposed 72,000+ images and 1M+ private messages through basic security failures and how proper Firebase rules and authorization could have prevented it all.

Built an AI assistant to automate my administrative tasks - repurposed a Discord bot with a Mastra agent, MCP server for tool access, and Mem0 memory layer.

Engineering insights on shipping fast without breaking things: how structured development processes turn reactive firefighting into proactive building.

Tech leadership isn't about having all the answers - it's about asking the right questions to help engineers think critically and solve problems independently.

My thoughts on why AI won't replace engineers, but engineers who use AI effectively will replace those who don't.

Exploring how local-first architecture can eliminate loading times and create delightful user experiences in modern web applications.