top of page

Our Story

John Harby

 

Founder & CEO, Autonomic AI, LLC

 

 

Overview

 

John Harby is the Founder and CEO of Autonomic AI, LLC, a research-driven technology company advancing next-generation AI architectures with mathematically grounded accuracy, equivalence, and energy efficiency. His work introduces the Accuracy AI framework, demonstrating significant reductions in computational cost and energy usage while maintaining verified correctness.

 

With university education focused on pure mathematics, John Harby entered the software industry circa 1990. He has held engineering roles at Oracle and BEA Systems and senior engineering and architectural roles at IBM, HP, Intuit, and Lion Bioscience, contributing to distributed systems, service-oriented architecture, and applied artificial intelligence. He is a co-author of the SOA Blueprints specification and has participated in multiple industry standards bodies, including the OASIS SOA Reference Model Technical Committee and several Java JSR expert groups.

 

He was quite fortunate to be associated with highly accomplished people such as:

 

  • Serge Lossa (database expert consultant who had worked with Peter Chen at one time)

  • Kurt Robson (Oracle - creator of the manufacturing data model)

  • Martin Fowler (Netscape consultant at the time)

  • Mark Carges (BEA - creator of Tuxedo at Bell Labs)

  • Rajiv Gupta (inventor of web services and GM at HP for e-speak) and his team

  • Mark Canales former VP of Lion Bioscience expert in Chemistry and an excellent teacher

  • Walter Gilbert (Nobel Laureate and visionary of the Lion Bioscience life sciences integration)

  • Steve Wilkes (lead author of The SOA Blueprints, successful CEO) 

  • many others particularly co-participants on expert groups and standards bodies, members of the  JAX-RS expert group are an example

​

His current research focuses on deterministic AI code generation, equivalence-class transformations, and energy-optimized computation, positioning his work at the intersection of mathematics, computer science, and scalable AI systems. Results include the invention of the Functor Model Architecture, a new approach to machine learning where the model learns via functional modification rather than parametric updates. Functor models are mathematically proven to reduce energy consumption by a minimum 78% versus their stochastic counterparts. An overall energy reduction for all AI operations of a minimum 80% is also proven. This is historic as when combined with other innovations within the industry we seem to have a good chance of controlling the AI energy problem. 

 

 

Professional Background

 

Harby’s career spans enterprise architecture, distributed systems, and applied AI across leading technology organizations:

 

  • HP (Smithsonian winning e-speak platform): Autonomous service composition and distributed systems

  • Oracle: Manufacturing and enterprise data models

  • IBM Global Services: Service-oriented architecture and enterprise integration

  • Intuit: Compiler and code generation systems for financial applications

  • Lion Bioscience: Neural modeling and bioinformatics integration

 

These experiences inform his current work in self-evolving AI systems and morphic model architectures.

 

 

Standards & Industry Contributions

 

Harby has contributed to multiple standards bodies and industry specifications, where designs are reviewed through multi-organization consensus:

 

 

He also presented at the OASIS Symposium (2007, San Diego) on SOA status, ROA and future direction.

 

Research & Publications

 

Harby’s research focuses on the formal relationship between energy, accuracy, and equivalence in AI systems. His work is available through platforms including Google Scholar and Zenodo.

 

Key areas include:

 

  • Energy-optimized AI computation

  • Deterministic and verifiable code generation

  • Equivalence-class transformations and canonical representations

  • Morphic and functor-based model architectures

  • An agentic ecosystem

 

Selected works include:

 

  • Energy, Accuracy, and Equivalence: A Mathematical Framework for AI Optimization

  • Two-Level Morphic Reduction: Global Quotient Topology and Local Fiber-Wise Transform

  • Homology and Galois Framework for Neural Model Spaces

  • Ontology Alignment and Equivalence-Class Transformations

 

 

Intellectual Property

 

Harby is the inventor of multiple U.S. patent filings in applied artificial intelligence, including work on:

 

  • energy-efficient AI computation

  • accuracy-constrained code generation

  • adaptive event detection

  • multi-criteria model selection

 

These innovations form the foundation of Autonomic AI’s morphic and Six Sigma frameworks.

 

 

Current Focus

 

His current work explores:

 

  • functor-style AI architectures

  • equivalence-driven computation

  • reduction of redundant inference

  • edge and energy-efficient AI deployment

 

This research aims to redefine how AI systems are built, evaluated, and optimized at scale.

 

       Links

​

Google Scholar 

GitHub

LinkedIn

ORCID

Zenodo

Enterprise and research-focused software. No consumer data collection.

bottom of page