How To Use Agent Swarms For Continuous Red-Teaming | Webinar

​Attackers just got a new tool. DeepSeek released V4.1-Flash-1017 this month, a model built for coding and agentic work that scored 77.0 on the DeepSWE benchmark and competes with top proprietary systems from OpenAI and Anthropic at a fraction of the cost. Time-to-exploit is about to drop, and defenders need tools to catch up.

​Meanwhile, most companies still test their attack surface once or twice a year. Code ships daily. Assets spin up constantly. By the time a pentest report lands, it’s already out of date.

​Autonomous agent swarms are built for exactly this problem. They discover assets continuously, chain and validate real exploits instead of flagging noise, and in some cases ship fixes as pull requests for engineering teams.

​This talk covers what’s working today in AI-driven offensive security, and how to use AI agents to discover, validate, triage, and remediate real vulnerabilities.

Speaker Bio

​Brandon Veiseh is Co-founder and CEO of MindFort, a Y Combinator-backed company building fully autonomous AI security swarms that find, validate, and patch vulnerabilities in web applications. MindFort serves multiple Fortune 500 companies and has prevented thousands of cyberattacks with AI agents. Before founding MindFort, Brandon led product and engineering at ProjectDiscovery, where he led development of the Nuclei cloud platform from an open-source tool to an enterprise product. He then led the Generative AI group at NetSPI, building generative AI tools for human penetration testers.