AI observability as the engine of production-ready AI

Samuel Jaja, Generative AI Engineer at Capgemini UK, brings a talk grounded in real production work to Leeds Sharp on Thursday 21 May 2026. The full title is "From Black Box to Glass Box: Systems Thinking & Observability as the Engine of Production-Ready AI", and it does what it says: no toy examples, no slides full of architecture diagrams that never shipped.

The case study is a live multi-agent financial compliance system with five specialised agents covering OFAC sanctions screening, AML detection and regulatory rules enforcement. Every agent decision lands in an immutable Langfuse audit trail capturing token usage, latency, cost and data lineage in real time. Jaja walks through three architecturally distinct approaches to building the same system: raw SDK with custom parallel orchestration via asyncio; the OpenAI Agents SDK for rapid prototyping; and MCP (Model Context Protocol) for sharing tools across team and organisational boundaries. Each trades differently, and the talk is explicit about when to reach for which.

The session also covers Azure OpenAI specifically in the context of regulated industries — finance, energy, construction — where data sovereignty and auditability are requirements, not preferences. Attendees will leave with a clear mental model for when multi-agent architecture is genuinely warranted rather than fashionable, and why observability is the feedback loop that shortens build-debug-ship cycles rather than a monitoring concern bolted on afterwards.

Leeds Sharp is a .NET user group that has been running in Leeds for years. It meets monthly at venues around the city centre, draws a crowd of working engineers and is free to attend. Sponsors include Fruition Group, JetBrains and tronik. The group is hosted by JohnS and has a track record of bringing in speakers with real production experience rather than conference-circuit abstractions.

This particular event is at a new venue: Enablis at 1 York Place, LS1 2DR. Doors open at 6pm for pizza and drinks; the talk starts at 7pm. Some attendees head to the Editors Draft pub afterwards at 8pm.

Why attend

Jaja has shipped AI systems in regulated industries and founded PyData Hull, the youngest NumFOCUS chapter in Yorkshire. He spoke at DDD North 2026. This is not a vendor pitch or a "getting started with LLMs" overview — it is a working engineer explaining production decisions with real cost and latency data behind them.

The Langfuse audit trail, asyncio orchestration patterns and MCP tooling are all concrete and reusable. If you are building anything with agents or considering Azure OpenAI in a compliance-sensitive context, the architectural comparisons alone are worth the evening. Pizza is free. The pub is optional.

Agenda

• 18:00 — Doors open, pizza and drinks at Enablis
• 19:00 — Talk: "From Black Box to Glass Box: Systems Thinking & Observability as the Engine of Production-Ready AI" — Samuel Jaja
• 20:00 — Event ends; optional drinks at the Editors Draft pub (vote taken on the night)

Speakers

• Samuel Jaja — Generative AI Engineer at Capgemini UK; founder of PyData Hull and Atlasync AI Ltd; MSc AI and Data Science with Distinction, University of Hull; speaker at DDD North 2026.