I started Echology because I watched the same pattern repeat everywhere. Teams get access to powerful AI models and then throw raw, unstructured documents at them, because nobody showed them there's a structural step before reasoning begins.

That gap, between raw content and structured intelligence, is where I work. I build the architecture that connects documents to understanding: classification, decomposition, verification, and audit.

Everything I build runs locally. No cloud dependencies. No data leaving the building. For the industries I serve, that's not a preference, it's a requirement.

Founder

  • 13+ years in geotechnical and civil engineering
  • Ran a geotechnical engineering firm
  • Last role at LaBella Associates (ENR Top 200)
  • Founded echology, Inc. (Delaware C-Corp)
  • Building AI systems that reveal structure in business operations

What I believe

Five principles. Non-negotiable. They govern every line of code and every business decision.

  1. Truth over plausibility. If I don't know, I say so. No outputs that sound correct but aren't.
  2. Structure over volume. Less that is accurate beats more that is approximate.
  3. Provenance over assertion. Every claim traceable to its source. Every number cited.
  4. Perception over generation. Understand before responding. Read before writing. Classify before inferring.
  5. Integrity under pressure. No compromising accuracy for speed, convenience, or impressiveness.

What I think about

I pull from everywhere. Quantum physics and string theory for how fields propagate. Christian faith for why things exist in the first place. Greek and ancient civilizations for what patterns endure across millennia. Space weather, disaster cycles, and astrology for how systems move at scales we barely track. Simulation theory for what the whole thing might actually be.

None of these are separate to me. They're different frequencies of the same signal. The work is learning where to listen.

What I'm building

I founded Echology (Delaware C-Corp) to build AI systems that reveal structure in business operations. Local-first. No cloud dependencies. Deterministic preprocessing before probabilistic inference.

Two products came out of that work:

  • Signal Provenance -- file-level hash-chain integrity monitoring. $10,000/yr.
  • Sympatheia -- coupling detection from passive metrics. Free audit.

The core classification engine, Decompose, is open source on PyPI. Free. Deterministic. No LLM required.