JFrog, a software supply chain company, has officially announced the successful integration of JFrog Artifactory and MLflow, enabling a new machine learning lifecycle integration. MLflow is an open-source software platform initially developed by Databricks Inc. The launch of this new integration aims to provide JFrog users with a solution that allows them to easily build, manage, and deliver machine learning models, as well as generate AI-driven applications and other software development components in a simplified, end-to-end DevSecOps workflow. Through this integration, organizations can verify the security and source of machine learning models, ensuring responsible AI practices. This integration primarily addresses the following core issues: currently, up to 80% or more of machine learning models cannot be smoothly integrated into existing operations due to technical issues, resulting in frequent failures in creating new AI-driven applications. The integration between JFrog and MLflow combines MLflow's model development solution with mature DevOps workflows, providing organizations with end-to-end visibility, automation, control, and traceability to effectively overcome this problem. Yoav Landman, Chief Technology Officer of JFrog, said, "Through this integration, we aim to help organizations embrace and successfully deliver AI, as well as generate more AI-driven applications. One key aspect is enabling developers and data science teams to manage models reliably, just like managing other software packages. This includes using a universal, scalable, and single binary repository system to ensure version control, perform security checks, and manage the entire lifecycle of models." Building on the successful integration between JFrog and Amazon SageMaker and Qwak AI Ltd., the combination of JFrog Artifactory and MLflow provides even more convenience for machine learning engineers and developers using Python, Java, and R. They can now use Artifactory as a model registry and freely work with their preferred toolsets. Furthermore, the JFrog platform acts as a proxy for Hugging Face, allowing developers to easily access various open-source models while preventing malicious models and ensuring license compliance. The solution also integrates software security features and scanners provided by the JFrog platform to maintain risk-free machine learning applications. In the current context of rapidly increasing malicious models, the ability to detect such models has become particularly important. In February of this year, JFrog's security research team discovered hundreds of malicious AI models in the Hugging Face AI repository, posing a serious threat to data leakage or attacks. The integration between JFrog Artifactory and MLflow provides strong technical support in preventing such risks.