Using LLMs and Generative AI to Fix Software Vulnerabilities

Explore the groundbreaking potential of Large Language Models (LLMs) and generative AI in addressing software vulnerabilities during this keynote presentation. Delve into the challenges faced by large enterprises in managing complex systems, slow patching processes, and inadequate risk prioritization. Discover how cutting-edge AI models like StarCoder can be leveraged to create and validate secure, private, multi-file fixes that rival those of expert developers. Learn about the advantages of using open-source models for code generation and assistance, enabling enterprises to remediate vulnerabilities without relying on external APIs. Gain insights into empowering security and operations teams to move beyond monitoring and issue detection to actively fixing vulnerabilities, even without developer experience. Envision a future "DevLess Security" world where the burden of application security is no longer solely on developers. Presented by Asankhaya Sharma, Co-Founder and Chief Technology Officer of Patched.codes, this 57-minute talk offers a compelling look at the transformative role of AI in software security.