CFP Directory

As we hand more agency to machines, we’re creating identities that can act, but not always be held accountable. I research how to red team and secure autonomous AI systems before that gap becomes systemic risk. My work lives at the intersection of offensive security and the rapid, often untethered growth of artificial intelligence. As an AI/ML Researcher and Red Teamer, I don't just look for bugs; I map the boundaries of autonomous systems to ensure they remain resilient when the unexpected happens. From the intricate layers of LLM pipelines to the hidden vulnerabilities in blockchain and DevSecOps automations, I focus on uncovering risks before they become reality. At DEF CON 33, I had the opportunity to speak on the Policy Track about the legal frameworks for ethical hacking. To me, security is as much about the humans who defend the systems as it is about the code itself. Advocating for global safe harbor standards is a vital part of ensuring that researchers can continue to protect the digital world without fear. I believe that as we hand more agency to machines, our need for intentional, human-centered security only grows. Whether I am simulating a real-world attack on an AI-driven workflow or refining a policy for international safety, my goal is to provide a clean window of clarity in an increasingly complex threat landscape. I am always open to quiet conversations about the offensive side of security, the future of AI resilience, or the ongoing effort of building trust in technology.

He focuses on Java security and AI security. His work has helped many high-profile vendors, including Google, Amazon, Cloudera, IBM, Microsoft, and Oracle. He has presented at Black Hat Europe 2024, Zer0Con 2025, Off-by-One Con 2025, Black Hat USA 2025, DEFCON 33, and Zer0Con 2026.

Rafael Salema Marques (SWaNk) is an old-school VXer who tends to define himself as a malware enthusiast. He has been coding malware since the early 2000s.His contributions bridge the dark technical underground with rigorous scientific methodology. Over his career, he has published code and research in prestigious hacker e-zines including 29A and Phrack Magazine. Additionally, he already contributed to the LOLBAS project, discovering and documenting novel ways to weaponize native OS executables. Academically, he did his MSc research at the Instituto Tecnológico de Aeronáutica (ITA) focused on employing an artificial immune system approach to detect advanced rootkit activities and covert data exfiltration. Building on this foundation, his PhD research introduced a novel method for detecting highly evasive pivot attacks in complex networks. SWaNk is always available for coffee, beer, and malware projects.

## Investigation and Analysis of Malware Spreading Infection via MSI Package Files (BSides Tokyo 2023)Language: JapaneseAgenda: https://bsides.tokyo/2023/Slide (PDF): https://github.com/mopisec/bsides-tokyo-2023/blob/main/BSidesTokyo2023_MSIanalysis_dist.pdfRecording: Not Available## Reverse Engineering Malware Written in C++ with IDA and Semi-Automated Scripts (Training at GCC 2025 Taiwan)Language: EnglishAgenda: https://gcc.ac/archive/gcc_2025/Slide (PDF): Not AvailableRecording: Not Available## Inside Pandora's Box: dissecting the latest arsenal and tactics of APT27 (VB2025 Berlin)Language: EnglishAgenda: https://www.virusbulletin.com/conference/vb2025/abstracts/inside-pandoras-box-dissecting-latest-arsenal-and-tactics-apt27/Slide (PDF): https://www.virusbulletin.com/uploads/pdf/conference/vb2025/slides/Slides-Inside-Pandoras-box-dissecting-the-latest-arsenal-and-tactics-of-APT27.pdfRecording: https://www.youtube.com/watch?v=TUATnkPEsnc## Continuous Evolution of Tianwu’s Pangolin8RAT and Custom Cobalt Strike Beacon (JSAC2026)Language: EnglishAgenda: https://jsac.jpcert.or.jp/en/timetable.htmlSlide (PDF): https://jsac.jpcert.or.jp/archive/2026/pdf/JSAC2026_1_6_naoki_takayama_en.pdfRecording: TBA (Should be available on YouTube soon)