Technical posts about building and deploying production AI systems. Everything here is based on real projects I built and operate.
Most AI chatbots are FAQ scrapers with a text box. Here is how to build one that handles objections, tracks cart state, and actually closes deals — based on the production sales agent I built for LuxuriousComputers.com.
Running Qwen 27B on a Tesla M40 and RTX 3060 in a home lab. Real numbers on tokens/second, VRAM usage, and quantization tradeoffs — plus the gotchas nobody tells you about mixed-generation GPU setups.
How I deploy production AI backends across 10 live sites using Cloudflare Pages Functions, KV storage, and external APIs — with real patterns for lead capture, notifications, and rate limiting at $0/month infrastructure cost.