# Patrick Colm Audley

*Hacker · Full-Spectrum Technologist · Polymath*

A seasoned technologist with thirty years of international startup experience and a broad palette of scientific,
business and computing skills. I work the long boundary between deep technical research and pragmatic commercial
execution — whichever side a problem needs.

Currently I split my time between [Blackcat Informatics® Inc.](https://blackcatinformatics.ca/) — the consultancy
I've run since 1994 — and applied research at the intersection of Topological Data Analysis, manifold learning,
Neuro-symbolic artificial intelligence, and large-language-model semantics. The through-line across more than thirty
years has been the same: find elegant solutions to complex technical problems, exploit problem isomorphisms, and ship
code that does what it says on the tin.

Earlier in my career I was instrumental to numerous tech launches, acquisitions, pivots, and exits across telephony,
high-performance computing, crowdsourcing, bioinformatics, and security — over twenty-five startup launches across the
years. I've moved high-throughput systems to the cloud (a social platform supporting massive concurrent counts in the
early 2010s, scientific analysis pipelines for genetics and satellite imagery, multi-continent high-resiliency financial
deployments) and learned, every time, that minimum-footprint systems built on open-source primitives outlast the
elaborately-architected ones.

I'm an active member of the local Vancouver and Alberta startup and development communities and enjoy fostering new
talent. Dangle a good problem in front of me, tempt me with a cup of coffee, and I'll yak your ear off on just about any
topic.


Spouse: Erin Audley.

## Currently

- **Helping organisations shape their AI and large-data strategies** through the current technology inflection — not
the
buzzword version, the version where you have to actually decide what to build, what to wait on, and what to ignore.
- **Working at the convergence of energy, compute, and autonomous systems** — applying AI and systems engineering to
hard, commercially consequential problems in power generation, cooling, resilient infrastructure, and climate
technology.
- **Building the mechanics layer for neuro-symbolic AI** — combining higher-order semantic spaces, latent-space
dynamics, topological and manifold structure, and explicit logic. [Logit Loom](https://github.com/paudley/logit-loom)
makes token generation programmable and inspectable; [GMEOW](https://github.com/Blackcat-Informatics/gmeow-ontology)
supplies attributed claims, reasoning, revision, and portable memory.
- **Building [PurRDF](https://blackcatinformatics.ca/purrdf/) ** — the RDF 1.2 toolkit, now at 1.0. One from-scratch
Rust engine carried byte for byte into Python, WebAssembly/JavaScript, and C: native codecs, canonicalization, SPARQL
1.1/1.2, SHACL, ShEx, and entailment through Web Ontology Language 2 Description Logic. Ranked full-text search, spatial
predicates, and vector similarity are answered from SPARQL itself, in-process. The premise: a graph should mean the same
thing in every runtime, and it should not need a PostgreSQL instance standing beside it.
- **Maintaining a handful of small libraries** — the structured-error library [e](https://github.com/paudley/e) and a
fast C++ [LRU cache](https://github.com/paudley/lru_cache) first written in 2004 and still useful. Earlier Python
tooling (*pyqa_lint*, *git-ai-reporter*) has been folded into [coding-ethos](https://github.com/paudley/coding-ethos).


## Open Source

- **[PurRDF](https://github.com/Blackcat-Informatics/purrdf)** (Rust / Python / WebAssembly / C) — PurRDF is Blackcat
Informatics' RDF 1.2 toolkit, at 1.0 under full semantic versioning: one from-scratch Rust engine carried byte for byte
into Python, WebAssembly/JavaScript, and C. It provides an immutable, value-interned RDF 1.2 dataset with triple terms,
reifiers, and base-direction literals; native codecs for Turtle, TriG, N-Triples, N-Quads, RDF/XML, TriX, HexTuples,
JSON-LD, and YAML-LD; W3C RDFC-1.0 canonicalization; a SPARQL 1.1/1.2 parser, algebra, and governed evaluator; SHACL
Core, SHACL-SPARQL, and SHACL-AF validation; ShEx 2.1; entailment regimes from RDFS through OWL 2 RL to an OWL 2 DL
tableau, each returning a reasoning report; deterministic full-text search, float-free GeoSPARQL 1.1, and embedding
nearest-neighbour search answered from SPARQL in-process, so an RDF project no longer needs a PostgreSQL instance beside
its triple store; and the GTS single-file, content-addressed graph transport. Behavior is gated by the W3C SPARQL,
SHACL, shexTest, OWL 2, and RDFC-1.0 suites and by frozen cross-language GTS conformance vectors.
- **[Logit Loom](https://github.com/paudley/logit-loom)** (Rust) — Rust toolkit for observing, transforming, steering,
stopping, resuming, and accounting for token generation. Logit Loom provides backend-neutral contracts for ordered logit
transforms, exact-byte token observers, bounded generation controls, content-bound plans and receipts, plus a llama.cpp
adapter with identity-bound checkpoints and explicitly scoped steering.
- **[gmeow](https://github.com/paudley/gmeow)** (Python) — Local Gmail intelligence for agents and automation.
Loopback REST + MCP over Gmail, backed by PostgreSQL, semantic search, attachment sidecars, local object storage, and a
knowledge graph. Built for trusted single-user local systems rather than hosted mailbox ingestion.
- **[gmeow-ontology](https://github.com/Blackcat-Informatics/gmeow-ontology)** (Rust / Python / RDF 1.2) — GMEOW —
the Global Metadata and Entity Ontology for the Web — is an RDF 1.2-native ontology engine for machine and human
minds. It treats LLM outputs, human assertions, observations, and institutional facts as attributed, revisable claims
with provenance, confidence, evidence, temporal scope, standpoints, reasoning traces, and defeaters. Its flagship use is
grounded agent memory: store, recall, and revise without collapsing disagreement or deleting audit history. Version
0.2.0 ships one engine as three products: a native Rust gmeow CLI, a grounded-memory MCP server, and a
content-addressed, append-only GTS ai-package with Python, Rust, Go, and TypeScript engines. Canonical self-contained
slices project into OWL, SHACL, JSON-LD, schema.org, citation, and Crossref deposit forms and align across the broader
linked-data ecosystem.
- **[gmeow-gts](https://github.com/Blackcat-Informatics/gmeow-gts)** (Rust / Python / Go / TypeScript) — GTS — Graph
Transport Substrate. A single-file, content-addressed, append-only transport for RDF 1.2 graphs and the binaries they
reference. GTS files are CBOR Sequence logs with deterministic fold semantics, BLAKE3 frame chains, optional COSE
signatures/encryption, suppression-based revision, and cross-engine conformance across Rust, Python, Go, and TypeScript.
It is the primary distribution method for GMEOW, but it is ontology-independent and does not require GMEOW.
- **[coding-ethos](https://github.com/paudley/coding-ethos)** (Go) — Policy-as-code enforcement for AI coding agents.
MCP server + CEL policies + git hooks + SARIF + static-analysis guardrails so an autonomous agent can't ship code that
fails your team's standards.
- **[ai-notes](https://github.com/paudley/ai-notes)** (Shell) — Working notebook for local-model experimentation, ML
papers worth tracking, and small reusable bits of inference plumbing. Public so the notes serve as their own changelog.
- **[nonlinear-semantic-graphs](https://github.com/paudley/nonlinear-semantic-graphs)** (Python) — Discovering
emergent knowledge-graph relations in high-order semantic vector spaces — manifold learning + spectral analysis
applied to LLM embeddings to surface relations that aren't explicit in the source corpus.
- **[lru_cache](https://github.com/paudley/lru_cache)** (C++) — Fast, thread-safe Least-Recently-Used cache as a C++
template, with a comprehensive unit-test suite. First written in 2004 and still useful — the kind of small library
that just needs to work.


## Publications

- **Emergent Knowledge Graphs from High-Order Semantic Spaces** (2025-10-01). Working paper · Blackcat Informatics Inc.
(2025).
- **Evaluating Usability and Security with Users of All Ages** (2007-04-30). CHI 2007 — Workshop on Usability and
Security.
- **OXBench: A benchmark for evaluation of protein multiple sequence alignment accuracy** (2003-10-10). BMC
Bioinformatics 4:47 (2003). doi:10.1186/1471-2105-4-47.


## Coverage

- **[Cogneto: to identity and beyond!](https://www.infoworld.com/article/2199876/cogneto-to-identity-and-beyond.html)**
— InfoWorld, 2007-05-25. InfoWorld article on Cogneto's identity technology; used as third-party support for Patrick
Audley's Founder and CTO role at Cogneto Development Inc. Supports: Patrick Audley was Founder and CTO of Cogneto
Development Inc. during the 2006-2007 adaptive-authentication period.
- **["Powered by Your Mind" Academics, Psychology, Technology Drive New Authentication
Solution](https://www.cutimes.com/2007/01/16/powered-by-your-mind-academics-psychology-technology-drive-new-authentication-solution/)**
— Credit Union Times, 2007-01-16. Credit Union Times article on Cogneto's cognitive-biometric authentication work;
used as third-party support for Patrick Audley's Cogneto role and adaptive-authentication work. Supports: Patrick Audley
worked on Cogneto's cognitive-biometric and adaptive-authentication technology as Founder and CTO.


## Patents

- **Globally Aware Authentication System**. US 2008/0222706 A1, published 2008-09-11.
<https://patents.google.com/patent/US20080222706A1>
- **System and method for user polling over a network**. US 8,510,385 B2, published 2013-08-13.
<https://patents.google.com/patent/US8510385B2>


## The Arts

Outside the technology work, I'm involved in [Odd Numbered Productions](https://oddnumberedproductions.ca), our arts
vehicle, where I make work at the intersection of AI, sensor data, and cryptographic art.

## Contact

- Email: <paudley@blackcat.ca>
- Phone: +1 (780) 670-0484
- ENS: `patrickaudley.com`
- Ethereum: <https://etherscan.io/address/0x3FEFC1B6D9b6A851bfC274fEa1929c31c6F23cf2>
- GitHub: <https://github.com/paudley>
- GitLab: <https://gitlab.com/paudley>
- LinkedIn: <https://www.linkedin.com/in/patrickaudley/>
- Crunchbase: <https://www.crunchbase.com/person/patrick-audley>
- ORCID: <https://orcid.org/0000-0003-4382-7625>
- Google Scholar: <https://scholar.google.com/citations?user=YgbTpqwAAAAJ&hl=en&oi=ao>
- X: <https://x.com/paudley>
- Twitter: <https://twitter.com/paudley>
- YouTube: <https://www.youtube.com/@patrickaudley>
- Stack Overflow: <https://stackoverflow.com/users/733505/paudley>
- PyPI: <https://pypi.org/user/paudley/>
- Facebook: <https://www.facebook.com/paudley>
- Reddit: <https://www.reddit.com/user/paudley/>
- Mail care of: Blackcat Informatics® Inc., 112 Westbourne Rd., Spruce Grove, AB, Canada T7X 1R6

---

Copyright © 2003–2026 Patrick Colm Audley. Original written content licensed
[Creative Commons BY-NC-SA CAv2.5](https://creativecommons.org/licenses/by-nc-sa/2.5/ca/). Generated
2026-09-03T11:27:55-06:00 (deploy 40a26ac41bdd097852c5d40608204f0ac8c0ea8a64ad084de4ff2481b8606f60).
