Writing
Deep dives on the engineering behind SaaS, AI, and web products — architecture, debugging, and lessons learned shipping to production.
How to build a real estate CRM in 2026 — the buyer-requirement data model, a site-visit-centered pipeline, portal lead capture, and buyer-to-listing matching, from building a real estate portal myself.
When custom CRM development beats Salesforce, and how to build one that lasts — the data model, pipeline state machine, idempotent lead capture, and MVP scope, from running lead pipelines for 5,000+ dealerships.
A practical, measurement-driven guide to Next.js performance — the levers I actually use: Core Web Vitals, shipping less JavaScript, images, fonts, caching, and third-party scripts, with links to the deep dives.
A practical guide to Generative Engine Optimization (GEO): how to write an llms.txt, allow AI crawlers, and structure content to get cited by ChatGPT, Perplexity, and Gemini.
A practical, end-to-end guide to building a RAG chatbot with Next.js, Postgres + pgvector, and an LLM — chunking, retrieval, streaming, and the production concerns (cost, caching, guardrails) most tutorials skip.
Add semantic search and RAG to your app with pgvector inside the Postgres you already run — and a clear framework for when a dedicated vector database like Pinecone is actually worth it.
A practical, first-hand guide to improving First Contentful Paint (FCP) in a Next.js / React app — measuring it, shipping less JavaScript, taming fonts, deferring scripts, and serving static HTML.
The four caches in the Next.js App Router — Request Memoization, Data Cache, Full Route Cache, and Router Cache — explained with code, plus a simple mental model for revalidation.