Offensive AI Security Researcher | Agent Systems & Exploitation

Founding AI Security Researcher (R&D)

We’re looking for a deeply technical, offensive-minded researcher to join as a founding member and shape the security layer of next-generation autonomous AI systems.

This is not a traditional security role — it’s about breaking, understanding, and rebuilding complex agent-based systems in real-world environments.

The Mindset

Offense-Driven Thinking

You instinctively analyze systems by how they can fail. When you see an autonomous workflow, you immediately identify weak points — reasoning flaws, memory vulnerabilities, and execution risks.

Builder + Breaker

You don’t stop at research. You take full ownership — from identifying vulnerabilities, to building proof-of-concepts, to implementing real-time protections in production.

Execution-Oriented

You are focused on outcomes. You ship working solutions that protect systems — not just theoretical insights or research papers.

What You’ll Do

• Design and execute advanced attack scenarios against autonomous agents (e.g., goal manipulation, memory corruption, infinite execution loops)
• Develop real-time guardrails that intercept and prevent unsafe or malicious actions before execution (API/tool-level control)
• Build and integrate security layers directly into production systems
• Create isolated, containerized environments to continuously test and validate agent behavior under real-world conditions
• Work closely with engineering teams to translate vulnerabilities into robust, scalable defenses

Requirements

• 5+ years of experience in offensive security, systems research, or security engineering
• Strong hands-on approach — from research to production implementation
• Deep understanding of system internals, exploitation techniques, and attack surfaces
• Ability to work across multiple layers — from low-level systems to application-level logic

Advantage

• Experience in advanced system exploitation or security research environments

Tech Stack

• Programming: Python, Go, Rust, Bash, C/C++, JavaScript/TypeScript, Java
• Systems & Infrastructure: Kubernetes, Docker, cloud platforms (AWS/GCP/Azure)
• Low-level tooling: eBPF, WebAssembly (Wasm), low-latency proxy architectures
• Isolation & runtime security: gVisor, Kata Containers
• AI agent ecosystems: LangGraph, CrewAI, AutoGen, OpenAI Swarm, n8n, Semantic Kernel

You should not only be familiar with these technologies — you should understand how they operate under the hood, and how to exploit their weaknesses.

What Makes You Stand Out

• You think like an attacker and build like an engineer
• You move fast from idea → exploit → solution
• You thrive in ambiguity and early-stage environments
• You bring both depth and ownership