About
Paul Boucherat
Operations management graduate. Two years from first build to deployed products across four industries.
Background
BA (Hons) Business & Operations Management from Nottingham Trent University, where I studied Goldratt's Theory of Constraints under Prof. Roy Stratton and Prof. John Disney. That education taught me to see systems differently: where bottlenecks form, why they move, what breaks when you optimise the wrong part.
Before tech: John Lewis, where I was selected for the Management Development Programme, youngest and only part-time candidate nationally. Then Stone Computers (2012–2016) as Service Buyer and Desktop Product Manager. International supplier management, CEO-level reporting, procurement, tender responses. Reduced “Jobs on Hold” from 13% to 2.5%.
I taught myself to code. No CS degree, no bootcamp. Python first, then TypeScript, then infrastructure. Two years ago I started building with AI, not as a novelty but as the thing that finally let me build what I could design.
ADHD (diagnosed), dyslexia (diagnosed), autism (pending assessment). It shapes everything: pattern recognition, hyperfocus, systems thinking. The neurodivergent advantage isn't a marketing angle; it's the reason the framework exists.
The combination is unusual: operations theory, commercial experience, self-taught engineering, AI-native development. I build systems the way production lines run. Find the constraint, fix it, measure whether it worked.
Philosophy
Everyone works differently with AI. Some people think in conversations. Others think in code. Some need structure first; others need to experiment until structure emerges.
I don't prescribe a method. I help people find their own AI working style, through their actual work, with their actual constraints. People who take control of their own AI learning always outperform the ones following someone else's playbook.
“He does not advertise himself. He does not fill a room with noise. What he does is think.”
“Paul’s patience is not passive. It is a specific skill.”
What I Build
End-to-end: APIs, MCP servers, AI agents, trading platforms, evidence analysis systems, web applications. All of it sits where AI meets domain expertise: property, trading, legal, infrastructure.
Technical Profile
Languages & Frameworks
Python, TypeScript, FastAPI, Astro, Streamlit, React
AI & Agents
CrewAI, PydanticAI, OpenAI API, Claude API, MCP (Model Context Protocol)
Infrastructure
Fly.io, Cloudflare Pages, Supabase, Docker
Domain Expertise
UK property data (Land Registry, EPC, Rightmove), trading (MT5, PineScript, spread betting), construction (Primavera P6), legal evidence analysis