#55: Dmitri Shifrin: Cybersecurity & AI at NVIDIA

We sat down with a System Security Architecture Manager at NVIDIA, to discuss the convergence of Automotive cybersecurity and AI, as well as NVIDIA’s Morpheus Cybersecurity Framework.

About Dmitri Shifrin

Dmitri is a System Security Architecture Manager at NVIDIA, leading a team of cybersecurity experts focusing on securing NVIDIA networking platforms, automotive networking solutions, and AI. With previous key roles in cybersecurity architecture at HARMAN and SanDisk, Dmitri brings extensive experience across multiple security domains, from hardware and embedded systems to cloud and AI applications. Throughout his career, he has spearheaded numerous customer engagements and collaborations in automotive, cloud, and AI security, showcasing a blend of technical expertise, business acumen, and leadership skills.

Summary of conversation with Dmitri Shifrin

In this episode of “Left to Our Own Devices, the Product Security Podcast,” Dmitry Shifrin, System Security Architecture Manager at NVIDIA, shares his extensive career journey and insights into the convergence of automotive, cybersecurity, and AI. Here are the key highlights and insights from the conversation:

Career Journey:

Early Career at SanDisk: Dmitry began his cybersecurity career at SanDisk, focusing on secure SSDs. This role provided hands-on experience in embedded security.
Transition to Automotive Security: Seeking greater challenges, Dmitry joined Towersec, later acquired by Harman, where he worked on intrusion detection systems for vehicles and integrated product security practices into Harman’s systems.
Role at NVIDIA: Dmitry now leads a team securing NVIDIA’s networking platforms, with a significant focus on AI infrastructure. His work spans hardware security, cloud security, and more.

Key Topics Discussed:

Automotive, Cybersecurity, and AI Convergence: Dmitry discusses the complexities of securing connected vehicles and the challenges posed by the integration of AI in automotive systems.
AI Infrastructure Security: He highlights the importance of securing the entire AI infrastructure, from chip level to cloud, to ensure robust and reliable AI applications.
Polymorphic Malware and AI: Dmitry talks about the rise of polymorphic malware, which can change its code on the fly, making traditional detection methods less effective.
Democratization of Development: AI is lowering the barriers to creating both beneficial applications and malicious software, requiring new approaches to threat modeling and cybersecurity.

Insights:

Technological Impact of AI: Dmitry emphasizes AI’s potential to revolutionize industries like healthcare and education, bringing unprecedented capabilities and opportunities.
AI and Security: The conversation underscores the dual role of AI in cybersecurity – both as a tool for enhancing defenses and as a means for adversaries to create sophisticated attacks.
Securing AI Models: Ensuring the security of AI models involves protecting data throughout its lifecycle and addressing vulnerabilities at every stage of the AI training and deployment process.
Future Challenges: Dmitry notes the increasing complexity of cybersecurity in the AI era, with the need to secure vast amounts of data and manage the risks associated with AI-enhanced attacks.

Memorable Moments:

Professional Achievement: Dmitry highlights the announcement of NVIDIA’s Blackwell platform, which his team worked on, as a significant professional milestone.
Personal Fulfillment: He shares a rewarding experience of lecturing students from peripheral areas in Israel, inspiring the next generation of tech professionals.

Dmitry’s insights provide a comprehensive view of the evolving landscape of cybersecurity in the context of AI and automotive technologies, offering valuable lessons for both industry professionals and aspiring students.