Engineer. Consultant. Builder.

I’m a Lead AI Engineer at AWTG Ltd in London, building production-grade AI systems for real-world teams. My work spans RAG and document-grounded workflows, GraphRAG-style knowledge systems, agent-style orchestration, evaluation, response validation, guardrails and cloud-native delivery on Azure AI Foundry and GCP — deliberately at the intersection of consulting, product thinking and hands-on engineering.

At AWTG I contribute to the KAI AI platform — production LLM workflows, Python/FastAPI services, data pipelines and reusable AI capabilities used by 65,000+ active users across major organisations, including the British Council. I’ve led AI engineering delivery for an education platform (not yet public) and built optimisation and decision-support capabilities for telecom networks. Each of those has a case study on the Work page.

What I enjoy most is translating ambiguous business problems into reliable technical systems. A vague “we should do something with AI” becomes a defined problem, a data-readiness check, a small honest prototype, an evaluation harness and — eventually — a system that people trust because it demonstrably works and knows its own limits.

I care about the unglamorous parts of AI delivery: document quality, retrieval evaluation, guardrails, cost, observability and the governance conversations that make a system defensible. Those are the parts that decide whether AI survives contact with an enterprise.

I write about all of it in Reliable AI, my independent newsletter on building production-ready AI agents, RAG systems, evaluation and enterprise AI delivery — practical, technical and free of vendor noise.