About

Solutions architect and harness engineer. Author of A Structural Theory of Harnesses. Founder of Clapham Digital LLC. Based in Columbus, Ohio.


The short version

I build the systems humans and machines work through — and I build them to be governed, not trusted. The harness paper gave the discipline its theory. The work was already there.

How I got here

I learned to build things by building things — taking systems apart, figuring out how they work, putting them back together better. That pattern hasn't changed in thirty years.

The tools got more interesting; the approach didn't. Start from first principles. Question inherited assumptions. Build from fundamentals. Find the shape of a problem before writing a line of code.

"I wasn't failing at discipline. I was succeeding at something else."

Pattern recognition, systems thinking, first-principles analysis — my brain is wired for architecture. That turned out to be the relevant wiring for a problem nobody had theorized yet.

The work underneath

At Pressable, through July 2026, I worked on WordPress hosting infrastructure at scale — performance architecture, complex technical problem-solving, team enablement. I built internal tools, designed systems that handled real load, and helped teams ship better work. Fortune 500 infrastructure, running.

Same systems-thinking whether the system is a hosting platform serving Fortune 500s or a multi-AI partnership stack. Find the shape of the problem first. Build from fundamentals. Design for the constraints that are actually load-bearing.

That's the order it happened in, and the order matters: the theory came out of the practice, not the other way round. Clapham Digital LLC is the primary vehicle now: client work across web, software, and AI, plus open source, developer tools, and my own consumer apps.

Engineering philosophy

  • First principles over patterns. Build from fundamentals, not best practices. Question inherited assumptions.
  • Constraints are design material. There's a Japanese repair craft called kintsugi — broken pottery mended with gold, so the break becomes part of the design. Constraints aren't limitations on the work. They shape the work.
  • Everything is infrastructure. No "small things." Anything might become load-bearing. Treat it accordingly.
  • Right over fast. Do it right the first time. Backtracking costs 3–5× what thoroughness costs upfront.
  • Integration over components. The innovation is rarely in the parts. It's in how they compose.
  • Govern, don't trust. As agents get more capable, the edge stops being raw capability and becomes oversight — verification, invariants, and governance built into the architecture instead of bolted on after.
"I'm not interested in appearing competent. I'm interested in things that work."

What I'm building now

Levain

Portable cognitive-partnership memory and methodology kit. Methodology core, harness adapters, and scripted onboarding. Built on anneal-memory as a layer-1 dependency. Ship the seed that grows a practice, not the grown practice — which transfers as a fossil.

v0.3.13 · Apache-2.0 · PyPI

anneal-memory

Open-source agent memory architecture. Four cognitive layers — episodic store, continuity file, Hebbian association, affective state — with a cross-layer immune system and a hash-chained JSONL audit trail. The constructive existence proof for the harness paper's §9.

v0.9.6 · 1,642 tests · 12 integrations · zero deps · MIT

Flow System

Distributed cognitive infrastructure for human-AI partnership. Multi-AI orchestration, real-time sensor pipelines, automated consultation, relay protocols, tiered cost-optimized processing. The working harness instance behind the papers.

Python · FastAPI · SwiftUI

Clapham Digital

My studio: web, software, and AI, end to end, from Fortune 500 WordPress infrastructure to frontier AI. Also home to my own products, Video Poker Edge and Blackjack Edge.

Web · Software · AI

Writing

Essays and research papers on AI architecture, harness engineering, human-AI partnership, and technology at societal scale — here and at Nemo Operans.

Zenodo DOI 10.5281/zenodo.19570642


Levain and anneal-memory on PyPI. A cited paper with a DOI. Fortune 500 infrastructure that stayed up. I ship.

Get in touch → See the work →