Selected Work
A few selected builds, shown with the same restraint as the work itself: what it is, what I made, and the stack behind it.
A production-grade lakehouse built on GCS and BigQuery, turning raw flight telemetry into cost-aware, decision-ready models with fully automated ingestion.
A local-first, AI-native operating system for personal knowledge work — replacing hidden chaos with reliable workflows, multi-agent orchestration, and sub-second audit logging for persistent memory.
Tools & Technologies
I'm self-taught in data and AI — no formal role, just genuine curiosity and a lot of hands-on building. My best work sits at the intersection of data engineering, automation, and decision support, where the goal is not just to move data around, but to create something a person — or a business — can actually operate, trust, and improve.
A lot of modern tooling creates the illusion of progress while increasing hidden chaos. I care about the opposite: clear structure, reliable workflows, and systems that stay understandable as they grow. That means I lean toward local-first thinking, explicit protocols, human-in-the-loop AI, and practical delivery over inflated platform ideas.
What drives me is that I don't think in isolated pages, dashboards, or scripts — I think in systems, even while I'm still learning the ropes. I look for the operational bottleneck, the data boundary, the handoff problem, and the simplest structure that can carry real work. Whether I am building a pipeline, an automation flow, or a knowledge system, the aim is the same: make the machine useful without making the human disappear.
[ Place ryan-portrait.jpg Here ]
Personal Pulse
Solving vertical puzzles. I enjoy the intersection of physical power and technical problem-solving on the wall.
Whether it's sport or trad, I value the focus and resilience required for long multi-pitch routes.
Serious about the craft. From precision grinding to dialled-in espresso shots. It's a system in itself.
Thinker. Builder.