AI Exchange - Nova Intelligence, WeBuild-AI, AWS
AI EXCHANGE LONDON
Marcello Urbani
Nova Intelligence
Adel Rahimi
WeBuild-AI
Elliott Almeida
Amazon
Join us on Wednesday 6th May for our next AI Exchange event, hosted by AWS at their Holborn HQ.
Here's what's on the agenda:
Nova Intelligence are tackling a problem that doesn't get talked about enough, most AI agents simply weren't built to work with enterprise software like SAP. Their talk discusses how they have solved that, and what it looks like when AI works as well in the enterprise world as it does everywhere else.
WeBuild-AI are making the case that bigger isn't always better. Their talk breaks down how small language models, through distillation, fine tuning and low rank adaptation, can deliver the same results as larger models, at a fraction of the cost.
AWS are tackling the question underneath all the others, not how to use AI, but what it's doing to the human mind. Their talk explores how cognition is shifting, which human traits become more valuable as AI takes on complexity, and what genuine human AI partnership actually looks like.
As always, doors open at 6:30pm and talks start at 7pm sharp. We have a tight capacity for this event, so please RSVP as soon as you can to secure your space.
Agenda:
Marcello Urbani @ Nova Intelligence
"Bringing Agentic AI to the SAP Ecosystem"
Most enterprise software wasn't built with AI agents in mind — and mainstream agents weren't built for enterprise software. SAP is the sharpest example: code lives in the system, not a filesystem, and git-based workflows simply don't apply.
Nova closes that gap — an agentic assistant that speaks SAP natively. This talk covers the mismatch problem, how we designed Nova around SAP's environment, and what it looks like when an AI agent finally works as well on SAP as it does on mainstream systems.
Bio: Marcello is a software engineer with over two decades in the SAP ecosystem, most of them building DevOps tooling at Basis Technologies. Last year he joined Nova Intelligence to focus on AI applied to SAP systems. An active open source contributor, he is best known for abapfs, a VS Code plugin that enables ABAP development directly in the editor.
Adel Rahimi @ WeBuild-AI
"Small Language Models: Practical AI for the Real World"
Large language models have dominated the AI conversation, but a quieter revolution is underway. In this talk, we'll explore the rise of small language models and why they matter now more than ever, from reduced infrastructure costs to data sovereignty and the ability to run AI on-premise or at the edge. We'll dive into the techniques that make small models punch above their weight, including distillation, quantisation, fine-tuning and low-rank adaptation. Whether you're navigating regulatory constraints, working with limited compute, or simply looking for a more practical path to production AI, this session will give you the tools to think smaller and smarter.
Bio: Adel is a Principal Engineer at Webuild-AI. He has a decade of experience in building and scaling AI and ML solutions at different industries and different company sizes, ranging from series A to unicorns and more recently he was a Sr. AI Engineering Manager at Procter & Gamble. He has also published several research papers in the areas of AI Explainability and NLP and is an advocate of safe and explainable AI.
Elliott Almeida @ AWS
"The Cognitive Partnership: How AI is Reshaping Human Intelligence"
This session explores the profound psychological transformation at the intersection of human cognition and artificial intelligence. Moving beyond tactical implementation, we examine how AI is fundamentally altering our cognitive identity and which human psychological profiles naturally thrive in AI collaboration. Drawing on the "Missing Middle" philosophy and the tension between cognitive offloading and scaffolding, we investigate the paradox of augmentation: as AI handles increasingly complex cognitive tasks, what uniquely human capacities become more valuable?
We hope to see you there for another great AI filled evening!
