Open architectures you can build from.
How I'd design common AI systems — retrieval over long or sensitive documents, agents, guardrails — written up in enough detail to use. Free to use and adapt; a link back is appreciated.
- Blueprint
- My own design for a problem, detailed enough to build from.
- Teardown
- My take on someone else's published method — what works, what breaks, how I'd adapt it.
- Note
- A shorter opinion or lesson learned.
All blueprints
- Blueprint
Chunking strategies, compared
How you cut documents into pieces decides what your AI can find. Seven ways to do it, what each costs, and the order I try them in.
Built & tested23 Sept 2026 · 5 min - Blueprint
RAG for long documents: give it a map
Ordinary RAG answers 'find me the paragraph' well and 'what does this 400-page file say overall' badly. Keeping the document's structure, with summaries at each level, lets one system answer both.
Built & tested23 Sept 2026 · 4 min - Blueprint
Agent memory without replaying the whole conversation
Most agents remember by pasting the entire history into every request, which gets slow, expensive and eventually stops fitting. This design splits memory into tiers, built from open-source parts, so each turn carries only what it needs.
Built & tested23 Sept 2026 · 5 min - Blueprint
Open-source OCR: route pages, don't pick a model
Open OCR models now rival paid APIs, but the best one for a scanned form is wasted on a clean PDF. This pipeline sends each page to the cheapest tool that reads it well and escalates only when it can't.
Built & tested23 Sept 2026 · 4 min - Teardown
RAG patterns: which ones are worth it
There are dozens of named RAG techniques and most teams need four or five. What each pattern does, what the evidence says, and the order I'd reach for them.
Built & tested23 Sept 2026 · 5 min
Contact
Working on something
like this?
I'm open to AI engineering, architecture and training work. Tell me what you're building and what the constraints are — that's usually enough to start.
- Taking on new projects
- Usually replies within a day