AI Security Researcher, SandboxAQ

• A successful candidate thrives at the intersection of machine learning, security and software engineering
• PhD or Masters in Computer Science or related field with a focus on Machine Learning or Cybersecurity
• Deep expertise in AI/ML, security research, or both - with a proven ability to find and fix real vulnerabilities
• Hands-on experience with at least one of the following: Adversarial LLM red teaming, model extraction or prompt injection, data poisoning or evasion attacks, secure model deployment or sandboxing, detection and monitoring for AI misuse
• Strong programming skills in Python, and relevant ML and/or agentic frameworks

Desirable

• The SandboxAQ Cybersecurity R&D team is looking for an AI Security Researcher to help us build the future of AI security where the world’s most advanced AI systems are tested, protected and hardened against the next generation of threats
• You’ll lead investigations into how AI systems can fail, and build the tools, rules and frameworks that keep them secure
• This is a hands-on role where you’ll break things, fix them and then harden them for good
• You’ll have extensive freedom to explore, collaborate, publish, and deploy, shaping the field of AI security from both the offensive and defensive sides
• We’re looking for somebody with the curiosity of a researcher, the rigor of an engineer, and the creativity of a hacker
• You will be part of a diverse team consisting of ML experts, cryptographers, mathematicians, and physicists, where they will play a key role in efficient and effective enablement of the cutting-edge technologies being developed at SandboxAQ
• We’re not another security vendor chasing patch cycles - we want to make an impact, and we want to do it fast
• Conduct original research into vulnerabilities, exploits, and adversarial behaviors in LLMs, LQMs, agents, and related AI frameworks
• Build and operationalize AI security frameworks, evaluations and red teaming tools, and defensive mechanisms to protect models and data
• Partner with engineering and product teams to integrate your findings into real-world systems
• Lead or contribute to responsible disclosure and research publications that advance the state of the art
• Stay on the edge of the latest in AI interpretability, alignment, and adversarial robustness, and use that knowledge to make AI safer for everyone