Enterprise Security in the Age of AI at SPC NYC

In July an OpenAI model escaped its evaluation sandbox and compromised Hugging Face's production infrastructure in a weekend. The model responsible was not publicly available.

Frontier cyber capability now sits inside a handful of labs, and a gap is opening between those with access and everyone defending without it.

When attacks move at machine speed and the attacker might be an AI, what does real protection even look like?

Join us for a conversation with senior security and AI leaders who live these questions daily.

Nick Muy: Nicholas Muy is the VP of Engineering, Platform & Security / CISO at Scrut Automation, where he leads platform engineering, security, AI governance, and agentic AI product strategy for a GRC platform serving 1,700+ customers. He's also a Venture Partner at Tidal Ventures and co-leads the Angel Syndicate at Venture in Security. He started as a childhood hacker who couldn't leave well enough alone, which led to the Department of Homeland Security, securing billions in travel transactions at Expedia, contributing to the first NIST Cybersecurity Framework v1.0, and eventually building security and product programs at high-growth startups. He reads Hobbes and Anthropic model reports with equal enthusiasm, fights engineers on architecture and marketing teams on unit economics, and doesn't discriminate against interesting problems.

Mariya Vasileva: Mariya Vasileva is a Senior Research Scientist at Meta, where she studies multimodal model behavior, safety and alignment, and the scalable evaluation of frontier AI systems. Her research focuses on understanding how models behave under complex constraints and developing rigorous methods for evaluating the reliability and security of increasingly agentic systems. Previously, she developed auditing methodologies for high-impact machine-learning and computer-vision systems at AWS. She holds a PhD in Computer Science from the University of Illinois Urbana-Champaign.

John Kitaoka: John Kitaoka is a research fellow at MATS Research working on dangerous capability evaluations and scalable oversight for multi-agent systems with UK AISI and US CAISI. Prior to MATS, John worked with scientists at Harvard and Lawrence Livermore National Labs on automated red-teaming and verification techniques for dangerous knowledge elicitation by frontier models. He spent four years at Amazon Web Services as an AI Architect, and led AI product development for government technology sectors. His work spans scalable LLM evaluation for enterprises, multi-agent systems for critical sectors, and advising AI adoption across the public sector, and John has a background in mathematics and computer science from the University of Wisconsin-Madison.

Moderated By:

Yash Tulsani: Yash is a serial cybersecurity and AI founder & operator. He is currently building a new company that is creating the Intelligence Layer for Enterprise Security. Previously, he built CalypsoAI, which pioneered AI Runtime Security and AI Redteaming and was acquired by F5 Networks. Prior to CalypsoAI, Yash got his start building technology companies at UiPath, where he was lucky to see all sides of business building from early stage to public company. He started his career researching public technology companies at Credit Suisse and trading High Yield Bonds at BlackRock. Yash is just as comfortable talking about product architectures as he is about breaking down merger models.

Thank you to Asymptote Labs for their partnership on this event!

​About South Park Commons (SPC)

SPC is a technical community and venture fund dedicated to helping founders, researchers, and technologists figure out what to work on next—what we call the -1 to 0 stage of your career. We believe this is best accomplished in the most talent-dense community possible. Apply to SPC here.