I have spent more than 40 years building, testing, managing, and improving software and systems—from BASIC, dBase, FoxPro, and SQL to modern web applications, local AI, agent workflows, and RAG/GraphRAG.
I am building a Personal Life Operating System: a local-first, human-governed environment for AI agents, retrieval, knowledge graphs, automation, and personal information. The private system remains private. The public Personal Life OS Core shares a sanitized, MIT-licensed foundation that others can inspect and reuse.
I want the public work to be useful to someone who has not seen my private system. These three projects give you concrete methods and examples to inspect:
- PLOS v2.1 brings documents, source-backed research, workflows and review into one application. Start with the research walkthrough and the fictional nightly examples. Installation and AI routes need your own configuration.
- Photo Knowledge shows how reviewed names, face regions, captions and keywords can travel with a picture through standard metadata. Try the small XMP sidecar generator or inspect the embedded JPEG examples. The examples use fictional labels; the tool does not identify people.
- Cyber Review Workbench runs a small local security-check demonstration, saves action receipts and compares the result with a clean control. Python's standard library is enough. A passing synthetic check leaves the human review open; it does not certify a system.
- Professional overview and résumé
- Explore Personal Life OS Core
- Read the latest release
- Watch: I Built a Personal Life Operating System—Here Is What It Actually Controls
- Watch: From Apple II to AI Agents — 5 Lessons That Still Matter
- 13 pull requests merged into Colibri as of August 25, 2026
- PLOS v2.1 with inspectable source, synthetic examples and documented review boundaries
- Practical work across software development, databases, quality assurance, infrastructure, cybersecurity, and technology leadership
- Local-first when personal information is involved
- Human approval for consequential actions
- Evidence, audit trails, rollback, and reproducible tests
- Upstream contribution and lawful open-source reuse
- Useful AI on ordinary consumer hardware
If this work has helped you, you may leave an optional one-time tip. Tips help offset hardware, storage, electricity, hosting, and testing. A tip does not purchase support, consulting, access, priority, or any deliverable.