Dewank

Agentic AI systems combine autonomy with real-world tool use, offering transformative potential but also introducing novel risks. This talk will introduce agentic AI and outline key vulnerabilities, including jailbreaks, tool hijacking, model misuse, and indirect exploitation through prompt injection. It will then focus on security risks associated with the Model Context Protocol (MCP) and conclude with mitigation strategies and practical defenses to enhance the safety of autonomous AI systems in real-world deployments.

Biography:

Mrinal

Generative AI has captured the imagination of students, researchers, and entrepreneurs alike, yet while prototypes are easy to build, very few cross the “last mile” into production systems that deliver real-world impact. That last mile—where performance, trust, and scale converge—is often the hardest, and it is rarely taught in classrooms. This keynote looks beyond the buzzwords to reveal the untold lessons of taking GenAI from proof-of-concept to enterprise-ready deployment. Drawing on real industry experiences, we will explore why the real challenge is not building a chatbot but making it reliable, fast, and cost-efficient at scale; how invisible infrastructure like orchestration, observability, and optimization layers keep systems running under real-world constraints; and why trust—through hallucination management, prompt injection defense, and transparency—becomes the ultimate differentiator. We will also examine how the most successful deployments blend human expertise with AI automation, and the emerging practices such as efficient model fine tuning and serving that will shape the next generation of solutions. This talk is about the untaught lessons: what separates a clever prototype from a system that truly transforms industries. Attendees will walk away with an understanding of the hidden tradeoffs and design principles that define the future of enterprise-scale GenAI—and how to prepare themselves to solve these challenges as they enter the workforce.

Biography:

Mrinal Karvir is a Senior Cloud Software Engineering Manager at Intel, where she leads the Intel Developer Cloud for the Edge, helping innovators build and deploy AI on the latest Intel platforms. She has spearheaded breakthrough projects, including Intel’s first presence-aware PC experience, which won a CES Innovation Award. Recognized as an Ethical AI Champion, she shares her expertise at global forums to promote responsible and impactful AI. Mrinal also serves as Vice Chair of IEEE Women in Engineering in Santa Clara Valley, building strong networks for women in tech. She is passionate about mentoring and inspiring the next generation of engineers and AI leaders.

Dr. Vishnu S Pendyala

Email:

Address:One Washington Sq, San Jose State University, San Jose, United States, 95192-0250