We have pioneered the concept of “ML for the 99%” with the ML at reasonable scale series, and recently discussed what changed with the new AI wave (spoiler: not much, the fundamentals stay!). In the talk we review the basics of ML in production and stress what changed and what didn’t in the era of LLMs.

Generative AI and machine learning systems are reshaping industries but also introducing new security risks. The reliance on vast data, rapid deployment cycles, and automated pipelines in MLOps has expanded the attack surface, exposing vulnerabilities to data poisoning, adversarial inputs, and pipeline exploitation.

This session explores the unique security challenges of ML systems in the GenAI era and provides actionable strategies to safeguard them. Learn why traditional approaches fall short and how to fortify your ML lifecycle to
stay ahead in an evolving threat landscape.

An Ideal MLOps platform in the Generative Al era is a comprehensive solution that supports the entire machine learning lifecycle, fromdata preparation and model development to model deployment and monitoring. It should provide seamless integration of tools and technologies that enable organizations to build, deploy, and manage machine learning models with ease. In this talk, we'll review some of the tools and solutions that exist to build, deploy, and scale GenAl workloads across different environments successfully.

Discover how incorporating continuous human guidance ensures ethical, accurate, and context-aware outputs, while autonomous agentic systems push the boundaries of AI, enabling proactive decision-making and complex problem-solving. And see a demo of creating a governed chatbot on your own company’s data within 60 seconds – not 6 months!