Why I am the person building it
A human has an intent: run this safely, classify this risk, approve this transaction, build this application. Somewhere between intent and execution, meaning, control and trust get lost. I have engineered the same thread across industry after industry, and it has now arrived at the point I believe defines the next decade of enterprise software.
The conviction behind FluidIntent comes from the research underneath it — an MSc in Intelligent Systems and a PhD in robotics and neural networks, studying how biological systems commit to action under uncertainty. Reliable action under uncertainty depends on the quality of the model committed against, before the action is released. It is the same problem biological motor control solves, now appearing in agentic software.
Today, governed agentic action. Tomorrow, safe physical action, closing the loop back to the robotics work I started decades ago. Human intent in, trusted action out, without losing control.
Founded Merlin Systems in 1998 and ran it for fourteen years building bipedal humanoid robots. Sole inventor of the Servo Air Muscle, the Micro-Proportional Air Valve and the Absolute Optical Linear Position Sensor.
PhD in robotics and neural networks and an MSc in Intelligent Systems: machine learning, vision, SLAM, Monte Carlo localisation and stochastic methods, with a long thread of biologically inspired learning including novel Hebbian and agglomerative clustering methods.
Led the architecture of a real-time wind-energy control platform for RWE/E.ON, onboarding live data from over 10,000 turbines into 24/7 control rooms, turning a failing first version into a fault-tolerant second within twelve months.
Principal AI Developer at the UK Hydrographic Office, setting technical direction across multiple teams and building generative and agentic AI capability into a public-sector navigation body.
Distributed systems built from first principles: a CHORD key/value store, an MVCC eventually-consistent database and a serverless agentic and RAG framework. Graph and ontology work going back to early Bayesian risk analytics.