Field Notes
Patterns and mental models picked up building real products — from AI agents and LLM systems to drag-and-drop editors and production Go services. Not tutorials, but hard-won lessons and the reasoning behind them.
AI & LLM
02 notesBuilding a Provider-Agnostic LLM Gateway
One gateway, many providers — normalizing Claude, Gemini, and OpenAI behind a single request format with a pluggable codec layer, central cost tracking, and zero-redeploy model swaps.
Designing AI Agents that Use Tools
What turns an API call into an agent — the tool-calling loop, giving agents the right tools and context, grounding answers with retrieval, and pausing for human approval.
FRONTEND
03 notesReact Performance Patterns
Memoization, virtualization, code splitting, and avoiding unnecessary re-renders in large-scale React apps.
Building Without Libraries
Creating drag-and-drop, resizing, snap alignment, and layout editors using vanilla DOM events.
State Management at Scale
Redux Toolkit patterns, Context API tradeoffs, and when to use which approach.
BACKEND
04 notesGo for Web Services
Building production REST APIs with Fiber, clean architecture, and graceful shutdown.
Database Design & Optimization
Schema design, indexing strategies, query optimization, and choosing the right database.
Authentication & Authorization
JWT-based auth, RBAC implementation, permission matrices, and security best practices.
Real-Time Communication
WebSocket patterns, event-driven architectures, and scaling real-time features.