Hi, I'm Francis. I build AI systems enterprises and customers trust
Building cognitive products since 2016...
When it was called NLU
-
Are your pilot AI projects not delivering value in production?
-
Want AI that integrates with your business operations?
-
Need to engineer around LLM hallucinations?
-
Looking for auditable, explainable systems?
-
Tired of consultants who talk but can't build?
About me
I'm Francis, AI Engineer and Cognitive Architect based in Edinburgh. I work with enterprise executives and tech scaleup leaders who've experienced the gap between AI promises and production reality. Whether you're dealing with a failed AI implementation or trying to scale beyond proof-of-concept, the core challenge remains: building cognitive systems that deliver reliable business value.
My journey in natural language understanding and cognitive architectures spans a decade, starting in 2016 when language models were tiny and "AI" meant something very different. I've developed production systems at Market Logic Software, AstraZeneca, Yelp, Twitch, and Sage Group, focusing on systems that combine continuous space reasoning (LLMs) with discrete process control flow (business rules).
My approach centres on cognitive architectures, context engineering, agentic workflows, knowledge graphs, and conversational intelligence systems that orchestrate business workflows and glean insights from user intents, journeys, and decision patterns.
When I'm not building AI systems, I mentor AI engineers and advise business executives on building reliable AI products. I believe in empowering organizations to build AI they can trust through systems that are auditable, explainable, and actually work in production.
Why work with me?
Here's what sets me apart and how I can help drive value for your business:
-
Cognitive Architectures That Work
I build "reasoning on rails" systems where LLMs make structured decisions within deterministic workflows. This isn't prompt engineering or RAG tweaks. It's architecting type-2 reasoning systems that combine neural intelligence with symbolic control flow, creating automation you can trace, audit, and trust in production.
-
Production-Grade Evals & QA
Most AI projects optimize for demos instead of production quality. I've built comprehensive evaluation frameworks that measure what actually matters: answer accuracy, grounding, hallucination rates, and response quality. From development-time testing through CI/CD gates to production monitoring, these systems ensure your AI maintains quality at scale, not just in controlled demos.
-
Strategic Intelligence From Operations
Beyond fixing individual failures, I architect systems that extract business intelligence from user interactions. Conversation flows reveal user journeys, decision patterns, and friction points. What starts as customer support becomes strategic insight for product and business decisions.
-
Explainable by Design
Black boxes don't survive enterprise scrutiny. Every cognitive system I build includes structured decision traces, audit trails, and explainability tools. Not as an afterthought for compliance, but as core architecture that enables debugging, improvement, and trust.
-
Let's have a virtual coffee together!
Want to see if we're a match? Let's have a chat and find out. Schedule a free 30-minute strategy session to discuss your AI challenges and explore how we can work together.