Principal AI Security Researcher: LLM Eval & Post-Training

Palo Alto Networks, Inc.; Santa Clara (CA); On-site; USD 163,000 - 264,000; Full time

Job summary

Palo Alto Networks is seeking a Principal AI Researcher to advance cybersecurity capabilities in large language models and autonomous AI agents. You will craft training environments, data, and evaluators for tasks spanning vulnerability research, threat analysis, detection, and secure coding.

Cross-functional collaboration accelerates turning research into secure, production-ready features. Qualifications include deep cybersecurity expertise, experience with LLM-based systems, datasets, and

Qualifications

• Strong hands-on experience in cybersecurity areas such as vulnerability research, secure coding, threat detection, malware analysis, or incident investigation.
• Experience designing datasets, benchmarks, graders, and evaluation frameworks with verifiable outcomes.
• Experience with post-training methods like supervised fine-tuning, RL, or distillation.

Responsibilities

• Design reproducible training and evaluation environments for complex cybersecurity tasks.
• Transform code bases, vulnerabilities, security incidents, and detection logic into structured tasks for AI systems.
• Develop high-quality security datasets including synthetic data and adversarial examples.
• Build graders, verifiers, and reward signals for model evaluation and deployment.
• Collaborate with Security Researchers, ML Engineers, and product teams to move research into production.

Skills

Python

C/C++

LLMs

Security research

Code understanding

Education

BS/MS in CS/ML/AI/Cybersecurity

Tools

Docker

Job description

Palo Alto Networks is seeking a Principal AI Researcher to advance cybersecurity capabilities in large language models and autonomous AI agents. You will craft training environments, data, and evaluators for tasks spanning vulnerability research, threat analysis, detection, and secure coding.

Cross-functional collaboration accelerates turning research into secure, production-ready features. Qualifications include deep cybersecurity expertise, experience with LLM-based systems, datasets, and