Festo.IO_2023 - Automation and Optimization of ML Pipelines

Speaker: Samir Araújo de Souza (Amazon Web Services) ML model industrialization is a mandatory step for organizations that seek to scale up ML model building processes to hundreds or even thousands of training jobs per day. Automation is a natural path to improve time to market which allows you to deploy more ML models to production, minimizing failures, improving reproducibility/reliability, reducing cost/time to build new ML models, etc. You start by automating your ML Pipelines (MLOps), then you optimize the whole processes and finally you optimize each step for maximum performance. ML model optimization is an important topic in this process, given it can reduce the overall cost of your solution, improve user experience and be more sustainable. In this session, let’s explore both topics (Automation and ML Model optimization) and see how we can translate business challenges into “small industries” that build/deploy ML models to the cloud and to the edge. https://io.festo.com/