Securing Models: Safeguarding ML Systems in the GenAI Era

🎥 Recorded live at the MLOps World | GenAI Summit 2025 — Austin, TX (October 8, 2025) Session Title: Securing Models Speaker: Hudson Buzby, Solutions Architect, JFrog Talk Track: Latest MLOps Trends Abstract: Generative AI and machine learning models are transforming industries — but they also introduce serious new security risks. Model marketplaces like Hugging Face and Ollama are now filled with community models that often lack trusted authorship and may contain hidden vulnerabilities. In this session, Hudson Buzby explores the unique security challenges facing ML and GenAI systems today, sharing practical strategies for building and deploying open-source LLMs safely and at scale. He explains why traditional security frameworks fail to protect the ML lifecycle — and what steps organizations can take to build guardrails around their AI development pipelines. What you’ll learn: • Why AI models pose new and distinct security threats • How to identify and mitigate risks when using open-source LLMs • Why traditional DevSecOps methods fall short in AI workflows • How to design safe, scalable practices for AI model governance and validation.