My development settings
All about my development settings
A little space for
big ideas.
Notes on frontend, AI agents, and Web3.
Things I build, learn, and figure out along the way.
AI agents and harnesses
From tool loops to reliable systems.
How models, tools, and harnesses form an agent, with a bounded execution loop and practical extensions for context, permissions, recovery, evaluation, and coordination among agents.
Read the articlerepeat
Model. Choose the next action from the goal and available context.
Things I’ve been making
All projectsFresh from the notebook
All blogsAI agents and harnesses: from tool loops to reliable systems
How models, tools, and harnesses form an agent, with a bounded execution loop and practical extensions for context, permissions, recovery, evaluation, and coordination among agents.
Building enterprise design systems: from design patterns to frontend architecture
How to build and maintain an enterprise design system: shared patterns, component APIs, tokens, themes, documentation, and upgrades across product teams.
Demystifying modern AI: from LLMs to multi-agent systems
A comprehensive architectural breakdown of core modern AI building blocks: LLMs, workflows, agents, tools, MCP, skills, plugins, multi-agent systems, orchestrators, hooks, guardrails, and harnesses.
Stablecoins
What stablecoins are, who uses them and why, and how the stablecoins maintain their pegs
Web performance optimization
A deep dive into the key techniques for performance optimization: how to measure with Core Web Vitals and DevTools. Performance optimization with tree shaking, code splitting, and dynamic import.