AI Systems & Agent Workflows
Agent architecture, retrieval workflows, connected tools and explicit human checkpoints across several AI engineering projects.
Field work / personal project notes. Status: Multiple projects at different stages.
I have worked on more than one AI-related system. The common thread is not the number of agents or prompts. It is how a system keeps track of what it was asked to do, which source supports a decision, and where a person must still approve the next action.
A practical systems lab
My source records include AI systems engineering, an AI systems lab, knowledge retrieval diagnostics and skill/tool workflows. These are distinct strands of work with different maturity levels.
I care about bounded actions, dependable retrieval, version histories and stopping when evidence is missing. A convincing explanation does not replace a working integration or a permission check.
Don't confuse a diagram with a deployment
Some parts are architecture, others are experiments, and some involve real local tooling. I am not presenting all of them as one finished, publicly available AI platform.
This page is a window into a changing workbench: investigate the problem, keep the source, build carefully, verify behavior and leave a clear trail for the next iteration.