Blog

An honest comparison of Next.js and Vite — key differences, when to pick each, and a clear recommendation.

How I built a local LLM API server using Ollama that replaces OpenAI API calls with self-hosted models for zero-cost inference — architecture decisions, key challenges, and what I'd do differently.

How I built an automated blog generation pipeline using DeepSeek chat API with retry logic, content validation, and sitemap generation — architecture decisions, key challenges, and what I'd do differently.

How I built a custom MCP server that exposes internal tools and data sources to LLMs via the Model Context Protocol standard — architecture decisions, key challenges, and what I'd do differently.

How I built a multi-agent pipeline where a supervisor agent delegates tasks to specialized research, coder, and reviewer sub-agents — architecture decisions, key challenges, and what I'd do differently.

How I built a retrieval-augmented generation chatbot that queries a custom knowledge base and returns accurate, cited answers — architecture decisions, key challenges, and what I'd do differently.

How I built optimizing a real production website for performance, accessibility, and SEO — architecture decisions, key challenges, and what I'd do differently.

How I built a TODO Chrome extension using pure ES6 with no frameworks — architecture decisions, key challenges, and what I'd do differently.

How I built scraping real estate data while mimicking human behavior to avoid bot detection — architecture decisions, key challenges, and what I'd do differently.