Remote AI Scientist: Agentic Systems for Cybersecurity

CrowdStrike Inc.; Lincoln (NE); Remote; USD 120,000 - 180,000; Full time

Benefits offered by this job

Equity awards

Wellness programs

Vacation & holidays

Parental leave

Professional development

Employee networks

Office culture

Great Place to Work Certified

Job summary

CrowdStrike is seeking a Data Scientist to help build agentic systems for cybersecurity. You will train models, build AI agents, and rigorously measure performance on unique data streams generated from trillions of events daily.

You will collaborate with cybersecurity experts, design RL-based training, and prototype advanced agent architectures while contributing to production-ready solutions with a focus on reliability and impact.

Qualifications

• PhD-level depth of understanding in modern ML research; equivalent mastery acceptable
• Experience training generative models and LLMs
• Strong background in RLHF/RLAIF, PPO/GRPO/DPO and reward modeling
• Experience building agentic systems: ReAct, planning, retrieval/memory and tool use
• Fluency with GPUs, PyTorch and LLM stacks (Transformers/TRL/PEFT, DeepSpeed)
• Clean, reproducible research engineering and experimental tracking
• Ability to work independently on ambiguous objectives and communicate in large teams
• Proven ability to use AI to improve decision-making and workflows

Responsibilities

• Work at the intersection of AI and Threat Research
• Collaborate with cybersecurity subject-matter experts to understand analyst workflows
• Post-train LLMs and agents, fine-tune and reinforcement learn for tasks
• Develop agentic systems with planning, tool calling and memory modules
• Prototype state-of-the-art methods and evaluate with rigorous benchmarks
• Optimize prompts and inference for better model performance
• Coordinate with Eng, Data Science, and Managed Services to move prototypes to production
• Stay current with AI research to identify prioritizable areas

Skills

ML fundamentals

Probability & stats

LLM training

Reinforcement learning

Agent architectures

Prompt optimization

PyTorch/GPU

Research engineering

Independent work

Cybersecurity interest

Education

PhD-equivalent mastery

Tools

Transformers

PEFT

Deepspeed

vLLM

Job description

CrowdStrike is seeking a Data Scientist to help build agentic systems for cybersecurity. You will train models, build AI agents, and rigorously measure performance on unique data streams generated from trillions of events daily.

You will collaborate with cybersecurity experts, design RL-based training, and prototype advanced agent architectures while contributing to production-ready solutions with a focus on reliability and impact.