Archive
Notes
Short reflections and practical observations from projects, technical work, and commercialization problems. Treated as a working notebook rather than a publication.
AI systems
n8n vs. Claude Code (and other AI agents) for automation
When AI coding agents are the right tool, when a workflow orchestration platform is, and why the strongest pattern is usually hybrid.
Open note →LLM wikis and knowledge graphs: what should persist?
Why retrieval is only part of the picture, and how LLM wikis and knowledge graphs hold the understanding behind the documents.
Open note →Agent memory: when live state matters
A companion note on agent memory: why robotics, logistics, and other state-driven workflows need more than documents and a knowledge graph.
Open note →Building useful AI tools without turning everything into a chatbot
Notes on workflow design, data structure, retrieval, and where language models actually help.
Open note →
IP & commercialization
Enabling the black box: what does a biotech software patent need to teach?
An educational exploration of reproducibility, claim scope, and technical documentation in computational biotechnology.
Open note →How I think through IP strategy
A practical note on matching patents, licensing, trade secrets, FTO, and ownership questions to the way a technology may reach use.
Open note →
Engineering systems
Failure modes are part of the design
Failure modes, applicability domain, and model qualification, the same idea in two vocabularies, with a worked example from PBPK-driven dose optimization.
Open note →Optimization is not just maximizing a number
Reflections on tradeoffs, objective functions, constraints, sensitivity analysis, and why optimization depends on defining the right problem.
Open note →