Security Researcher

Research the security of AI systems end to end, from the agents and frameworks they're built on to the runtimes and infrastructure that run them
Take apart complex software and find where it breaks
Turn a vulnerability into a full, working exploit, not just a proof that the bug exists
Understand the defenses modern systems lean on, and how to get past them
Share what you find, through coordinated disclosure and conference talks

4+ years in security research, vulnerability research, or a closely related field
A strong sense of how software works under the hood, whether that's a modern framework or the system underneath it
Comfortable reading complex code, and reverse engineering with IDA Pro or Ghidra when there's no source
Working knowledge of common vulnerability classes, from unsafe deserialization and logic flaws to memory corruption, and the experience to take one all the way to a working exploit
Excited about AI, and eager to get to its security problems first
Real curiosity, and the drive to go deep when the answer isn't obvious

Nice to have

Experience researching AI systems, agent frameworks, or the infrastructure behind them
Published research, whether CVEs, blog posts, or conference talks
A degree in computer science, mathematics, or engineering, or the hands-on experience to match

See also