Sara Basson works at Google as Emerging Markets Accessibility and Disability Inclusion Lead, with the goal of increasing access to assistive technologies and advancing advocacy around accessibility in emerging markets like Africa and India. Her previous role at Google was  “Accessibility Evangelist,” focused on making the experience of Googlers more accessible and usable, through enhancements to technologies and education and advocacy around accessibility. In previous roles, she worked at IBM Research on Speech Technology, Accessibility, and Education Transformation.  Sara holds an M.B.A. in Management/Marketing from Stern School of Business, New York University, and a Ph.D. in Speech, Hearing, and Language Sciences from The Graduate Center of the City University of New York.  She also holds an Honorary Doctorate degree from Saint Mary’s University in Halifax, Nova Scotia, in recognition of her extensive activities using speech technologies to enhance accessibility in the educational arena. Sara serves on the Board of Directors for Lexington School for the Deaf and inABLE Africa.

Sara Basson works at Google as Emerging Markets Accessibility and Disability Inclusion Lead, with the goal of increasing access to assistive technologies and advancing advocacy around accessibility in emerging markets like Africa and India. Her previous role at Google was  “Accessibility Evangelist,” focused on making the experience of Googlers more accessible and usable, through enhancements to technologies and education and advocacy around accessibility. In previous roles, she worked at IBM Research on Speech Technology, Accessibility, and Education Transformation.  Sara holds an M.B.A. in Management/Marketing from Stern School of Business, New York University, and a Ph.D. in Speech, Hearing, and Language Sciences from The Graduate Center of the City University of New York.  She also holds an Honorary Doctorate degree from Saint Mary’s University in Halifax, Nova Scotia, in recognition of her extensive activities using speech technologies to enhance accessibility in the educational arena. Sara serves on the Board of Directors for Lexington School for the Deaf and inABLE Africa.

Reclaiming Control: The Case for Sovereign AI and Repatriation with Dell Automation and ClearML

As AI models grow in complexity and regulation tightens around data governance, organizations are rethinking the centralized cloud model. The next frontier is Sovereign AI - AI that is trained, deployed, and governed within the boundaries of trusted infrastructure.

In this session, we will present how ClearML and the Dell Automation Platform enable enterprises and governments to repatriate AI workloads from public clouds while maintaining full operational agility. We’ll learn how open, automated MLOps pipelines can deliver data sovereignty, compliance, and cost efficiency, without sacrificing scalability or innovation.

We’ll explore real-world examples of how AI workloads can be repatriated, automated, and orchestrated across edge, on-prem, and hybrid environments, and what this shift means for the future of AI infrastructure and trust

Ready for Prime Time? Evaluating AI Agents for Real-World Enterprise Deployment

AI agents have rapidly emerged as the next major paradigm shift in AI system design, promising greater autonomy, deeper reasoning, and richer interactions with complex environments. Yet their growing capabilities introduce a corresponding rise in evaluation challenges, far beyond those faced in traditional LLM assessment. Whereas validating an LLM for enterprise use already requires rigorous benchmarking, safety checks, and policy alignment, evaluating agents adds new layers of complexity related to multi-step reasoning, tool use, environmental interaction, and dynamic behavior over time.

In this talk, I will describe how the landscape of agent evaluation is developing and highlight several emerging research directions that aim to better assess agent quality. This includes approaches for understanding how agents behave across multi-step tasks, methods for identifying exactly where problems arise during an agent’s execution, and techniques that help developers diagnose issues and iteratively improve their systems.

I will also discuss practical methods for ensuring agent robustness in real-world enterprise environments. This involves evaluating how well agents handle adversarial interactions, maintain stable performance under challenging or unexpected conditions, and consistently follow enterprise policies and security constraints. Together, these considerations illuminate what it takes to prepare AI agents for reliable, secure, and policy-aligned deployment at enterprise scale.