ATLANTIS:
AI-driven Threat Localization ANd Triage Intelligent System

Team Atlanta is an award-winning group, with accolades from Pwn2Own, DEFCON CTF, and numerous academic conferences. Our team comprises experts from Georgia Tech, Samsung Research, KAIST, and POSTECH. We are dedicated to advancing next-generation security research using Large Language Models (LLMs) to pioneer Security-AGI for the future. In DARPA’s AI Cyber Challenge, we aim to revolutionize security paradigms by leveraging AI and LLMs in our cyber reasoning system Atlantis.

Check our AIxCC Team
Banner image
feature image

Atlantis, a next-generation automated bug finding & fixing system

Intelligent beyond what you’d expect from a machine.

  • AI-Driven: Large Language Models (LLMs) powered bug detection and fixing for real-world software.
  • End-to-End Automation: Manages the entire bug-finding and fixing process.
  • Smart Patching: Fixing bugs automatically with human-like reasoning.
  • Multi-Lang Support: Compatible with complex languages such as C/C++, Java, etc.
  • Kernel Support: Supports complex systems like the Linux kernel.
  • Self-Improving: Employs LLM-powered techniques that enhance over time.
  • Autonomous Knowledge Inference: Applies domain-specific knowledge without human intervention.
More about Atlantis