Скачать или смотреть Deploying ML Models at Scale

This month, Vishnu Vardhan will talk about building ML pipelines for your Pythonhon applications and how to dynamicallylly scale them to fit your needs.

Talk Description:

In the realm of large-scale machine learning (ML) deployments, the ability to dynamically scale resources to meet varying workloads is essential for optimizing performance and efficiency. This presentation delves into the strategies and technologies for deploying ML models at scale, with a specific focus on autoscaling using Ray.io and Kuberay.
We will start by introducing Ray.io, a powerful distributed computing framework designed to simplify the scaling of Python-based applications. Attendees will learn how Ray.io can be utilized to enhance the scalability and efficiency of ML model deployments, leveraging its robust scheduling and task execution capabilities.

Building on this, we will explore Kuberay, an open-source project that seamlessly integrates Ray.io with Kubernetes, enabling efficient autoscaling of ML workloads. We will discuss the architecture and features of Kuberay, demonstrating how it facilitates the deployment and management of large-scale ML models in production environments.
By the end of this presentation, attendees will have a comprehensive understanding of how these technologies can be used to optimize their ML operations and drive innovation in their organizations.

Links:
Meetup - https://www.meetup.com/orlandopython/...
Ray - https://www.ray.io
Kuberay - https://github.com/ray-project/kuberay
Dask - https://www.dask.org