Speaker
Description
Recent advances in the Scientific Python Ecosystem have opened new possibilities for High-Energy Physics (HEP), especially through its integration with ROOT, the backbone framework for data storage, analysis, and visualization. With tools such as NumPy, pandas, and Jupyter now interoperating seamlessly with ROOT, researchers can build flexible and efficient workflows. At the same time, Machine Learning (ML) is becoming essential in HEP, particularly with the High-Luminosity LHC expected to produce exascale datasets. To address these challenges, the MLaaS4HEP framework provides an experiment-independent, cloud-based solution that enables full ML pipelines—from reading ROOT files to model training and inference—via simple HTTP calls. Deployed on INFN Cloud with features such as OAuth2 authentication, XRootD proxy access, and TensorFlow-as-a-Service, this system demonstrates how scalable and distributed ML services can support the next generation of physics analyses.