In this episode of the Security Repo Podcast, the team dives into the OWASP Top 10 for Large Language Model Applications with special guest Talesh Seeparsan, an expert in cybersecurity and AI safety. Talesh shares insights into why a specialized top 10 for LLM vulnerabilities is essential, delves into unique challenges like system prompt leakage and AI supply chain risks, and provides practical advice for small companies navigating AI compliance. The conversation wraps up with reflections on security best practices, including collaboration and skepticism about industry norms.

With over a decade in cybersecurity, Talesh is a trusted expert in protecting enterprise applications. He has collaborated with the U.S. DoD, developed Appsec training for Adobe, and secured identities for multibillion-dollar firms. Today, he guides companies in building secure, trustworthy, generative AI and LLM applications and volunteers with the OWASP Foundation team working on the OWASP Top Ten for LLMs. His north star is the safe, performant adoption of frontier AI.

https://www.linkedin.com/in/talesh

https://genai.owasp.org/resource/owasp-top-10-for-llm-applications-2025/

https://owasp.org/www-project-top-10-for-large-language-model-applications/

Ver todos los episodios