Agentic Offensive Security Demo Night

​NOTE: This is an AI Tinkerers event, if you're interested in demoing please sign up and submit a proposal.

AI Tinkerers San Francisco is hosting an Offensive Security Demo Night at WorkOS in downtown San Francisco, an evening built around live technical demos of AI-powered offensive security work. Expect real systems, real code, and real execution, from autonomous pentesting agents to browser exploitation and security loops for AI-generated code.

​The format is simple and deliberate. No slide decks, no polished product pitches, and no filler. Presenters are expected to show working demos, walkthroughs of attacks, and the mechanics behind how their systems actually behave. If you care about what breaks, how it breaks, and what it takes to automate that process safely, this event is designed for you.

​Talks currently include agentic vulnerability patching, automated exploit loops for AI-generated code, browser hacking with LLMs, secret protection for agents, and open-source attack techniques. Engineers and researchers from XBOW, the event’s featured sponsor, will also be in the room to discuss autonomous offensive agents, including architecture decisions, failure modes, and the safety boundaries involved in deploying them against live targets.

​This should be a strong fit for security engineers, researchers, red teamers, and AI builders working on agents that interact with untrusted environments. You can expect a technically serious room and conversations that go beyond surface-level takes, especially during the demo Q&A and the time between sessions.

​We’ll be gathering at WorkOS on Market Street in San Francisco. If you want an evening centered on live exploits, applied research, and the practical edge of offensive security, join us for a night of demos that show the work instead of talking around it.