Machine Learning Models: Select, Train and Use
Machine learning can often feel like a complex black box, but mastering its practical application boils down to these fundamental phases: Select, Train, Use (and Evaluate). In this talk, we will try to demystify the end-to-end ML lifecycle. We will cover elements of strategic decision-making behind choosing the right algorithm for your specific data problem, the best practices for training and fine-tuning robust models, and the crucial steps for deploying those models into real-world applications. More precisely, we will discuss:
Select:How to evaluate trade-offs and match the right algorithm, from simple regressions to complex neural networks, to your problem.
Train: Essential techniques for preparing data, optimizing hyperparameters, and avoiding common pitfalls.
Use: Bridging the gap between a local environment and production, including deployment strategies and performance monitoring.
Whether you are a data scientist or a developer looking to integrate AI into your workflow, you need a roadmap on how to use ML and AI models for turning raw data into reliable predictions.
Lecture at NEMO2026
Date/Time: Wednesday, July 29, 2026 at 08:30