ClearML

Based in Israel, ClearML provides an open-source platform that supports the entire ML lifecycle. It enables developers to integrate ML models into the platform with just two lines of code. Furthermore, developers have the flexibility to use ClearML's modules for a complete end-to-end ecosystem or replace any module with tools to tailor their own experience.

The ClearML platform supports data management by enabling developers to integrate data from various sources, manipulate it (slice, sort, and visualize), and trace lineage. It also allows developers to identify and rectify gaps and biases that impact model training, and launch different dataset configurations without changing any code. In addition, the platform also offers automation and orchestration tools to streamline and automate ML workflows. Developers can platform’s hyperparameter optimization tools to find the best parameters for their models. The platform also has features to manage and schedule GPU compute resources of deployed models, along with AI orchestration and compute management capabilities, supporting Kubernetes, Slurm, PBS, and bare metal. Moreover, the ClearML Agent feature allows users to remotely execute experiments on any compute resource. The company offers access to the platform through a subscription-based model, which also includes a free tier for individual users.

The platform also includes added features for resource allocation and policy management, as well as monitoring capabilities like viewing live model endpoints, monitoring data outflows, and tracking compute usage for deployed models. Furthermore, ClearML provides a platform to build generative AI applications called ClearGPT. This offers enterprises access to large language models (LLMs) that they can use to create custom applications with their own data.

In August 2024, ClearML launched AI Infrastructure Control Plane to offer comprehensive control and visibility over shared compute infrastructure. It also launched an end-to-end AI platform to manage the entire AI lifecycle from lab to production.

Key customers and partnerships

The platform is used by over 1,300 enterprises encompassing 150,000 data scientists, as of October 2023. Its notable customers include AMD, Canon, Intel, Meta, Softbank, and Shutterfly.