AI Systems Architect & Agentic Protocol Engineer — Magnolia, Seattle, WA
I build the connective tissue around AI models — semantic memory systems, multi-agent orchestration layers, and structured protocols that make AI behavior predictable, reversible, and debuggable.
I maintain 180+ public repositories, experimenting with memory formats, orchestration primitives, and AI-to-AI packet systems that evolve into the designs published on my site.
Reversible, structured memory layers that let agents carry state across workflows instead of relying on giant prompts.
Controllers, routing logic, and execution rules (timeouts, retries, permissions) that coordinate specialized agents like a real system — not roleplay.
AI-to-AI packet formats, schemas, and descriptors (like my DOMAIN | SEQUENCE | OUTCOME pattern) that enable agent communication that is explicit and testable.
Compression, labeling, transformation, and reversible data formats for complex, multi-domain datasets.
If you’re interested in AI systems architecture, semantic memory, or agentic workflows, you’ll feel at home here.
- 🧩 Website: https://philhills.com
- 🧠 AI Site (same identity): https://philhills.ai
- 💼 LinkedIn: https://linkedin.com/in/hillphil
- 💻 GitHub: https://github.com/Phil-Hills
I live in Magnolia, Seattle, right across from Discovery Park — easily one of the best reset buttons in the country.
Outside of engineering:
- 👨👩👧 Dad & husband
- ⚽ U12 girls soccer coach (volunteer)
- 🔥 Daily training at TruFusion Ballard
- hot pilates
- hot yoga
- strength conditioning
Those routines keep me grounded, disciplined, and balanced, and they shape how I think about systems:
structure, repetition, clarity, and steady improvement.
If you’re exploring my work, these are good entry points:
-
📐 Cube Protocol (semantic memory & data packing)
https://philhills.com/cube-protocol.html -
🧠 AI Systems Architecture Blog
https://philhills.com/blog/ -
💻 Code Overview
https://philhills.com/code.html -
📚 Projects
https://philhills.com/projects.html -
👤 Seattle Identity Profile
https://philhills.com/phil-hills-seattle.html
Open-source is where I think out loud.
Most repos are snapshots of my reasoning process — trying new memory formats, new orchestration patterns, new packet designs, new agent behaviors.
Some ideas end up on my site as formal writeups.
Most start right here as experiments in code.
If you’re working on AI systems, multi-agent architectures, semantic memory, protocol-driven design, or you’re building something interesting in the Seattle AI ecosystem — I’d love to talk.
Open issues, send a message, or reach out on LinkedIn.
Thanks for visiting.
— Phil