I am cleaning the public surface so it can hold two things at once: career proof for people evaluating me quickly, and the doctrine/writing that explains why the proof has this shape.
Current work
The technical focus is still agent infrastructure and physical AI, but I am trying to phrase it from the ground up instead of from buzzwords down. The shared architecture is simple: preserve the artifact, stabilize object identity, expose typed relations, keep hard validators outside the model, and learn only the residual ambiguity.
That pattern covers a surprising amount of my work: mechanical fixtures, CAD/DFM automation, tactical behavior graphs, agent harnesses, resume-portfolio grounding, and the DXF-to-object/GNN plan.
Interests
- Physical AI and CAD-native representation
- Geometry-native evidence, stable editable objects, typed relation graphs, deterministic checks, and learned repair only where the geometry remains ambiguous.
- Agent infrastructure
- Thin harnesses, explicit state, provenance, eval loops, and interfaces that improve as base models improve.
- Cognitive security
- Calibrated action under shift: evidence-bearing inference, analyst-legible drilldown, and confidence that does not pretend context is stable.
- Writing and graph memory
- A synced career and thought space: short public surfaces, source-preserved proof, dense marginalia, and graph views for the archive underneath.
- Music as method
- Motif, variation, timing, error recovery, and the discipline of keeping expression inside a form sturdy enough to carry it.
Prose
I am aiming for a barbell: brief, humble, grounded surfaces on one side; mystical density and old strange essays on the other. The middle should be mostly links, evidence, and carefully named doors.
The aphorisms matter because they keep a living edge. The resumes matter because they keep the edge accountable. I want the website to hold both without letting either one impersonate the whole person.
Excited about designing simple affordances within complex environments.