Speaker "Miro Enev" Details Back



Interactive Introduction to RAPIDS with an End-to-End Machine Learning Workflow Demo


RAPIDS is an open-source software stack focused on accelerating data processing and machine learning using NVIDIA GPUs. In this talk we'll motivate the need for accelerating the modern data science process, describe the design principles which power RAPIDS, and look at the similarity with the RAPIDS API and popular data science libraries (pandas and scikit-learn). We'll then walk through an interactive visual demonstration of the RAPIDS API and use it to perform ETL, model building/training, and inference -- all while comparing the performance of multiple CPU cores vs a GPU (single and multi-GPU).


Sr. Solution Architect & Data Scientist @ NVIDIA focusing on Deep Learning.