Brussels / 3 & 4 February 2024


AI for Developers: Treating Open Source AI as a Function

More and more applications are being developed which consume Artificial Intelligence (AI). AI is used in applications as diverse as fraud detection, medical detection, sales, security, and many more. AI adoption has more than doubled since 2017 and companies are now promoting an AI first strategy with their products and services. But if a developer wants to consume AI in their applications, do they need to be an AI expert? The simple answer is No!

In this talk application developers will learn how they can treat AI the same as any other "function" where they give it input and it returns the expected output. This is made possible by open source AI frameworks that enable users to manage models through a set of developer friendly APIs. Discussion will revolve around open source frameworks like Ray, TGIS, Triton and TorchServe, to give an understanding of the tools available to help on the AI journey. There will be a demo of one of these runtimes in action by loading open source models from Hugging Face from different domains and querying the models through a web UI.

There are great open source tools out there that provide an abstraction layer for application developers where they can consume AI models through APIs independent of understanding the data form of the model. Walking away from this talk, programmers will have a good appreciation of how they can practically consume AI in their applications without being an AI master.


Photo of Martin Hickey Martin Hickey