Скачать или смотреть Wednesday Webinar - Building Real-time Inference AI Services with FIWARE

There is no point in training Machine Learning models if there is no plan to make them available for applications to use. In this webinar, we are going to learn how to deploy an ML model as a REST API and leverage FIWARE components to enable real-time AI inference. The use case covered involves getting data from different sources/organizations for real-time inference with a trained ML model.

This webinar is your chance to learn the end-to-end implementation of the real-time inference AI service: ML Training, Docker setup, creating a Python Package, and implementing the ML API.

Presenter: Rihab Feki (Machine Learning Engineer and Evangelist, FIWARE Foundation)
Chapter: Processing
Difficulty: ★★★
Audience: Developers

00:00 Introduction
00:17 Learning Goals
01:16 What is real-time interference
02:57 Use case presentation
04:58 Overall architecture
10:38 FIWARE generic enablers used
10:54 Subscriptions set up
11:08 Overview of the AI service
12:35 Dataset and training
20:04 Jupyter Notebooks are not ideal for production
20:44 Structure of the ML package
27:03 High-level overview of the ML API architecture
28:29 ML API source code
31:54 Implementation steps
33:06 Demo
40:26 Resources