AI Tinkerers San Francisco: Offensive Security Demo Night [AI Tinkerers - San Francisco]
ait-events://invite/cp_rC-akj-qOJA
live session
$ ait events connect
resolving chapter... ok
loading event packet... found
attendee interface... ready
You're invited
$ rsvp --attend

AI Tinkerers San Francisco: Offensive Security Demo Night

Jun
25
Thursday
Thursday, June 25th, 2026 5:30PM to 8:30PM (PDT)
Address Info
Available on RSVP acceptance

Start your RSVP

First, verify your email. Then you'll finish your registration.

Already verified or registered? Sign in
Attendees 99+ registered
Talks include reverse engineering supply chain attacks, building LLM exploit loops, and hacking Chromium with agents, and more. View Demos »

⚡ Live Exploits, Autonomous Pentesting, and Raw Craft

(Banner) A promotional banner for an 'AI Tinkerers - San Francisco' Offensive Security Meetup on June 25 at 5:30 PM, featuring a cyberpunk-style illustration of three people coding at retro computer terminals. Text: AI Tinkerers - San Francisco Offensive Security Meetup - June 25 05:30 PM AI TINKERERS - SAN FRANCISCO Cyberpunk, retro-futuristic, neon | Colors: #3b4998, #0d0d1a, #ff007f, #00ffcc Note: The image is a promotional banner designed to advertise an upcoming event, combining informational text overlays with a themed background illustration.

Join AI Tinkerers San Francisco and XBOW for an offensive security demo night. We are bringing together engineers, researchers, and security hackers to showcase live, technical show-and-tells of systems getting popped, innovative exploits, and autonomous pentesting agents.

This is a night of raw craft, code, and execution. We have an absolute ban on slide decks and product pitches. Presenters must show working code, live deployments, and actual execution.

Our featured sponsor, XBOW, is the front-runner in autonomous pentesting. Founded by GitHub Copilot creator Oege de Moor, XBOW uses AI agents to continuously find and exploit web application vulnerabilities at machine speed, recently demonstrating its capabilities by reaching the #1 spot on HackerOne using only black-box access. Engineers and researchers from their team will be in the room to discuss the architecture, failure modes, and safety boundaries of running autonomous offensive agents.


📅 Event Details

  • Date: Thursday, June 25th, 2026
  • Time: 5:30 PM – 8:30 PM PDT
  • Venue: Hosted in San Francisco, CA (Exact location provided upon approval)
  • Capacity: 150 Builders (Strictly curated)

⏱️ Format & Schedule

  • 5:30 PM: Doors open, food, drinks, and Science Fair setup
  • 6:00 PM: Brief introduction by AI Tinkerers and XBOW
  • 6:15 PM: Main Stage Builder Demos (5-minute live runs + Q&A)
  • 7:15 PM: Science Fair & Networking (Hands-on code walkthroughs and peer-to-peer technical deep dives)
  • 8:30 PM: Doors close

🥽 Who Should Attend?

Entry is selective by design to maintain a high-signal environment. Space is limited to 150 participants, and our San Francisco events consistently reach capacity and move to a waitlist quickly. We require demonstrable proof of hands-on work with AI or offensive security.

This event is curated for:

  • Security Engineers & Researchers building LLM-assisted vulnerability discovery pipelines.
  • AI Engineers & Founders orchestrating multi-step agentic workflows that interact with untrusted environments.
  • Offensive Hackers & Red Teamers interested in the technical limits of autonomous exploitation.

📢 Submit Your Demo Proposal

We want to see live, technical show-and-tells of things getting popped, innovative exploits, or autonomous pentesting hacks. Tell us what you plan to break or demonstrate live. No slides, no product pitches—just raw craft, code, and execution.

Submit Your Demo Proposal

🥽 Speakers

North Korean Open-Source Attacks

Rene Brandel

Rene Brandel

CEO @ Casco

XBOW: Browser Hacking Agents

Nick Gregory

Nick Gregory

AI Researcher @ Xbow

Shai-Hulud: Agentic Secret Protection

Allie Howe

Allie Howe

Member of Technical Staff @ Keycard

Securing LLM-Generated Code

Arjun Krishna

Arjun Krishna

Robotics / Software Engineer @ WATonomous

LLM Security Loop for PRs

George Pickett

George Pickett

Founder @ Stealth startup


🤝 Sponsors

Featured Sponsors

(Logo) A minimalist logo featuring the stylized word 'XBOW' in white, set against a dark background. Text: XBOW Colors: #ffffff, #050e14 Note: The image displays a clean, stylized brand name or wordmark on a solid dark background, which is characteristic of a modern logo.
XBOW is an autonomous offensive security company that automates penetration testing to discover and validate software vulnerabilities. Backed by Sequoia Capital, the platform uses AI agents to simulate targeted, multi-step attacks directly inside customer environments. Unlike traditional scanners, XBOW independently validates every finding to eliminate false positives and can safely execute blackbox testing against live websites.

(Logo) A minimalist white logo featuring a stylized geometric 'K' symbol made of circuit-like lines, followed by the text 'Keycard' on a solid black background. Text: Keycard Colors: #FFFFFF, #000000 Note: The image features a brand name accompanied by a stylized graphic emblem, which is the standard format of a corporate or product logo.
Keycard is an identity, access, and governance platform built for AI agents.

As agents and multi-agent workflows take on real work - writing code, calling APIs, querying databases - they’re running on over-scoped, long-lived credentials that weren’t designed for this. Keycard fixes that by issuing scoped, short-lived credentials per action, enforcing policy at every tool call, and logging the full identity chain in real time. Agents get exactly the access they need - nothing more, and nothing standing still.

Additional Sponsors

(Logo) A logo for 'casco' featuring a stylized white helmet icon overlaid with a vibrant green four-pointed star, set against a solid black background. Text: casco Colors: #000000, #FFFFFF, #00FF9D Note: The image displays a brand name alongside a stylized graphic icon, which is characteristic of a brand logo.
Casco is an autonomous security testing company that continuously probes web apps, APIs, infrastructure, and AI systems for vulnerabilities. The platform deploys AI agents to run targeted, multi-step attacks against customer environments, surfacing issues like RCE, broken authorization, and prompt injection. Unlike traditional pentesters who test once a year, Casco tests year-round and validates every finding with step-by-step reproduction to eliminate false positives, delivering clear reports on what to fix and why it matters.

Venue Sponsor

(Logo) The image shows the WorkOS logo, featuring a white stylized bracket symbol followed by the text 'WorkOS' on a solid black background. Text: WorkOS Colors: #FFFFFF, #000000 Note: The image displays a corporate brand mark consisting of a unique graphic symbol alongside the company name, which is the definition of a logo.
WorkOS WorkOS is the modern API platform that empowers developers to quickly build and ship complex enterprise features, making any SaaS application enterprise-ready. It provides a set of essential building blocks and APIs for seamless integration with cloud services like Okta, Azure AD, and G Suite. Developers can rapidly implement features such as Single Sign-On (SSO), Directory Sync, and Audit Logs, significantly accelerating time-to-market for enterprise-grade infrastructure.


📸 Experience AI Tinkerers

Event photos

A group of people are gathered around a long wooden dining table inside a modern rustic event space, while a man in the foreground stands to speak to the attendees.
AI Tinkerers - San Francisco
A group of people are gathered around a long wooden dining table filled with colorful dishes, salads, and floral centerpieces in a modern indoor setting.
AI Tinkerers - San Francisco
A group of people are seated along a long, elegantly set dining table decorated with floral centerpieces and filled with plates, glasses, and bowls of food.
AI Tinkerers - San Francisco
A group of people are seated along a long wooden dining table set with white plates, wine glasses, and vibrant floral centerpieces in an indoor setting.
AI Tinkerers - San Francisco
A large group of people are gathered around a long wooden dining table set with plates, food, and flowers in an indoor space with a brick wall.
AI Tinkerers - San Francisco
A black and white photograph of a young man with light hair giving a presentation at a podium, holding a microphone in front of an audience with a projection screen displaying a website behind him.
AI Tinkerers - San Francisco

Interested in supporting future events or hosting a technical deep dive? View our sponsorship opportunities.


AI Tinkerers San Francisco Stats

  • Attendees: This elite community of 7,317 technical professionals features a powerful distribution of skills: 48% specialize in agentic systems and multi-agent orchestration, 42% in advanced model engineering and PyTorch, and 38% in high-performance AI infrastructure and MLOps. Distinguished by a high density of venture-backed founders and researchers from top-tier labs like Google DeepMind and Meta, members actively collaborate on cutting-edge Model Context Protocol implementations and verifiable AI governance.
  • Companies Represented: Featuring tech giants like Google, Meta, Apple, and Amazon, alongside AI pioneers such as OpenAI, Google DeepMind, Anthropic, and xAI, and rising startups like Replicate, Wordware, Chroma, and Vercel, and more
  • Demos: 487 demos were submitted and 172 were accepted and presented. The most compelling themes covered agent-native product UX, real-time multimodal voice/vision experiences, and protocol-driven integrations (especially MCP). Technical exploration has also emphasized reliability via tracing/evals and verification, grounding via advanced retrieval/context strategies, and security-focused agent infrastructure including safe tool access, privacy controls, and human-in-the-loop patterns.
  • Testimonials:
    “This is meaningful work. We don’t all recognize it but we need community and social activity... As someone exclusively working at home by myself I really really appreciate AI tinkerers. I’d love to participate again.”

Ready for more?

Check out other posts from this blog.

View all posts

Contact Organizers

Questions? We're here to help.