Blog
Writing about the technologies I work with day to day, the practice of using AI as part of the development workflow, and how open source models are changing what's possible.
October 25, 2026 · 10 min read
How I use Gemma, running locally through Ollama, to draft, validate, and ship posts for this site: a small Node script, structured JSON output, a build gate, and one command to deploy to Cloud Run.
October 25, 2026 · 8 min read
How I produce short marketing videos for a roofing company on a single 16GB AMD GPU: Flux for stills, a local LLM for prompts, Wan 2.2 for image-to-video, local TTS for voiceover, and Kdenlive to cut it together.
October 25, 2026 · 8 min read
A design-level look at an AI phone receptionist I built for a roofing company: what it does on every call, how the system fits together, and the design decisions that make it fast, reliable, and affordable for a small business.
October 21, 2026 · 11 min read
How I built a speech-to-speech AI receptionist for a roofing company that screens out marketers, live-transfers real customers, books appointments, and pushes leads to the CRM, on a single Google Cloud VM.
August 10, 2026 · 3 min read
A practical breakdown of where AI tools like Claude earn their keep in a professional engineering workflow — and where they still need a human in the loop.
July 22, 2026 · 3 min read
From early fine-tuned research checkpoints to today's near-frontier open-weight models — how open source LLMs got competitive, and what that means for teams building on them.
June 15, 2026 · 3 min read
Practical lessons on event-driven architecture, drawn from building order management and resource tracking systems on AWS EventBridge and SQS.
May 3, 2026 · 3 min read
Why pairing a Next.js frontend with a Java/Spring Boot backend is a pragmatic, boring-in-a-good-way choice for systems that need to last.
March 28, 2026 · 3 min read
Notes from building a retrieval-augmented generation proof of concept on Azure OpenAI and vector search — what worked, and where the real engineering effort goes.
February 11, 2026 · 3 min read
A practitioner's comparison of AWS and Azure for cloud-native architecture, after years of shipping production systems on both.