Over the Edge; Migration, Optimization and Intelligence for IoT devices with Microsoft Azure

What does a mature IoT Solution look like? Is it enough to just connect up to the cloud and go? What about my devices that can't speak Cloudy? How do I understand and control what this costs me? What about the promised AI that lives in the Cloud??

In this session, Andy introduces and explores the role of Edge devices in a modern IoT architecture and can be made intelligent.

There will be demos using Azure IoT tools, code, Cloud resources and simulated IoT Edge devices.

Preferred duration 60 minutes

Machine Learning with ML.NET and Azure Data Lake

ML.NET is an open source, machine learning framework built in .NET and runs on Windows, Linux and macOS. It allows developers to integrate custom machine learning into their applications without any prior expertise in developing or tuning machine learning models. Enhance your .NET apps with sentiment analysis, price prediction, fraud detection and more using custom models built with ML.NET

In this Session, Andy will show not only the core of ML.NET but best practices around Azure Data Lake and data in general when using .NET

Business Outcomes from AI

Andy Cross, Director of Elastacloud, Microsoft Regional Director, Azure MVP and all round good guy, gives a session on how to successfully build or transform a business using AI technologies.

Over the last years, Elastacloud have delivered analytics projects to a variety of customers. The greatest challenges around AI are both technical and organisational. The existing landscape of process and strategy doesn't solve these challenges in combination, and the gap between causes friction and the failure of AI projects.

When modelling the outcome of actions that were informed by AI, possibly enacted by AI, the standard risk modelling approaches need to be transformed to include a factor that can change over time to represent the effectiveness of the AI solutions. Given that we should accept errors as part of the AI solution, and that errors are reinforcing of better future decisions, we need to project risk as a decreasing vector over time.

How to Lose Friends and Influence AIs

What if you no longer needed anyone else? Two years ago I discovered I could ship real products with no team, no code review and nobody else to blame: just me, with AI as my pair, my team and occasionally my worst enemy. So I gave up my Microsoft MVP and Regional Director awards, sold my consultancy, vanished from the speaking circuit and set out to see how far that would go. Some of what I shipped was genuinely brilliant. Some of it was an unmaintainable shambles I'm still rather embarrassed about, though I'd like to think the gradient runs in the right direction. And the title is only half a joke: one of the ventures in this talk failed, and my co-founder is now fighting in a war partly because it did. Some production outages you don't get to roll back. This is an honest post-mortem of two years spent solo with AI: where two decades of engineering instinct saved me and where it sabotaged me; what I learnt about judgement, loss and which corners one can safely cut; and, by the end, which parts of a team you can genuinely replace with a model, and which parts you will miss far more than you expect. No slides full of best practice, just real stories, real code and the occasional scar.

Target audience: developers, tech leads and engineering managers of any stack who are using (or resisting) AI-assisted development. Level: intermediate; no prior tooling knowledge needed, the lessons are about judgement rather than any specific product.

Format: 45-60 minute session, adaptable to either slot. Story-driven with real code and demos from actual shipped (and abandoned) products; no live-coding dependencies, so nothing to go wrong beyond my dignity. Standard AV only: HDMI and a clicker.

This will be the first public delivery of this talk [or: first conference delivery; a shorter version was road-tested at a London meetup in September 2026].

Speaker background: former Microsoft MVP and Regional Director; previously spoke at Techorama, TechEd, Build, GoogleX, Msft OSS Day, before stepping back from the circuit in 2024 to found and run product businesses. Based in London; happy to travel, no restrictions on recording.