Andy started work as an electronic design engineer, specializing in computer architecture design in the UK and Germany.
During the period he was designing a new main computer for the Tornado fighter, he relaxed by revisiting a childhood interest in the first Neural Networks and creating a simulation of some simple Neural Net algorithms in C.
Returning to the UK in 1985, he created a Windows-based Neural Network development system. The first such commercial system in the world and arguably the first data mining product.
After an interview with a journalist, a small article in the Telegraph, entitled "Brainwave needs cash", created a flood of investment offers and led to the creation of Neural Computer Sciences.
In the following years he expanded the range of algorithms available to include Genetic Algorithms, Fuzzy logic rule induction and an interest in time series prediction.
An introduction to the head of future products at Reuters led to work on financial time series prediction, and through new investors to the creation of Science in Finance limited, arguably one of the first London FinTech companies.
Reuters lawyers were rightly cynical about black-box neural network models, so Andy switched focus to Fuzzy Logic Rule Induction as a machine learning algorithm.
A burgeoning interest in Genetic Programming and Chaos Theory led to Darwin, a futures trading-system generator using simulated evolution of a large set of fuzzy logic based trading models, and work on a PhD on the intersection of Chaos theory and machine learning.
Science in Finance was acquired by a Bermuda-based futures trading organization to get access to Darwin. The PhD work resulted in Andy's thesis "Time series prediction using supervised learning and tools from chaos theory".
After several years with Science in Finance, Andy founded and became CEO of Scientio, inc. with a New York investor, a publicly traded US company, marketing his fuzzy logic based software. ScientioBot was his first ChatBot, running on the now long forgotten Microsoft Messenger App.
Andy then diversified into NLP, creating Concept Strings, a mechanism for representing and recognizing sequences of concepts in text, that was acquired by a leading UK RegTech company.
Returning to his Fuzzy logic roots, Andy has created a large body of IP, which is slowly being released, centered round his rule language DARL, with IP to machine learn in supervised, unsupervised, reinforcement and association forms, to detect lacuna in rule sets, a ChatBot framework and much more to come.