Lead Agentic AI Security Scientist

CrowdStrike; United States; On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Market leader compensation

Wellness programs

Vacation and holidays

Parental leaves

Job summary

CrowdStrike is seeking an applied research lead to steer the agentic AI frontier in cybersecurity. You will set scientific direction, lead research on efficient, secure models, and develop rigorous evaluation methodologies across post-training, harness design, and agent verification.

The role blends hands-on experimentation with leadership, collaborating with security researchers, threat intel, and product teams to ground research in real-world operator workflows and adversarial conditions.

Qualifications

• MS or PhD in computer science, computer engineering, or similar quantitative field.
• 10+ years in data science, ML/DL, or AI with large-scale data focus.
• Hands-on expertise in modern LLM post-training techniques.
• Demonstrated research contributions to agentic systems.
• Strong evaluation discipline and rigorous methodologies.

Responsibilities

• Lead the research agenda for agentic AI systems across cybersecurity use cases.
• Drive original research on efficient models tuned for security workflows.
• Establish rigorous evaluation, verification, and validation methodologies.
• Partner with security researchers, threat intel, detection engineering, and product.
• Define abstractions for the skills and tools agents use.
• Advance agent design, harnesses, planning, memory, and multi-step reasoning.

Skills

Agentic AI research

LLM techniques

Evaluation methods

Leadership

Education

MS/PhD in computer science / engineering

Job description

CrowdStrike is seeking an applied research lead to steer the agentic AI frontier in cybersecurity. You will set scientific direction, lead research on efficient, secure models, and develop rigorous evaluation methodologies across post-training, harness design, and agent verification.

The role blends hands-on experimentation with leadership, collaborating with security researchers, threat intel, and product teams to ground research in real-world operator workflows and adversarial conditions.