[Remote] Data Scientist, Agentic Systems (Remote)

Note: The job is a remote job and is open to candidates in USA. CrowdStrike is a global leader in cybersecurity, dedicated to stopping breaches and redefining modern security. They are seeking a Data Scientist to build advanced agentic systems, leveraging unique cybersecurity data to train models and develop AI agents that enhance security operations.ResponsibilitiesWork at the intersection of Artificial Intelligence and Threat ResearchWork closely with subject-matter experts in cybersecurity to understand analyst workflows and their security operations proceduresPost-train LLMs and agents — supervised fine-tuning and reinforcement learning (RLHF/RLAIF, PPO/GRPO/DPO, reward modeling) — to automate analyst procedures and improve reliability on real security tasksDevise AI agents and combine them into increasingly complex workflows: planning and reasoning loops, tool and function calling, and retrieval and memoryResearch new approaches to agentic planning, and prototype state-of-the-art methods from the literatureEstablish objective criteria for benchmarking agentic systems — evals, LLM-as-judge pipelines, and trajectory-level metrics, with real statistical rigorOptimize prompts and inference to get the most out of every modelCollaborate and coordinate across Engineering, Data Science, and Managed Services teams, and partner with engineers to take prototypes toward productionKeep track of developments in the field of Artificial Intelligence and help identify, define, and prioritize areas for researchSkillsExcellent foundations in machine learning, probability, and statistics, with sound instincts for uncertainty, statistical skew/variance, and experimental designPhD-level depth of understanding in modern machine learning research —a doctorate itself is not required, but we expect equivalent mastery, including the ability to read, critique, implement, and improve upon current papersExperience training generative models, with a strong command of LLM training fundamentals (architecture, optimization, tokenization, data, and scaling behavior)Reinforcement learning / post-training as a core skill: RLHF/RLAIF, policy optimization (PPO/GRPO/DPO), reward modeling, and building RL environments for agentsExperience building agentic systems: agent architectures (ReAct, planning, reflection), tool and function calling, and retrieval/memory/context managementExperience with systematic prompt optimization, and with designing and building evals for LLM systemsFluency with GPUs, PyTorch, and the common LLM training and serving stack (e.g., Hugging Face Transformers/TRL/PEFT, DeepSpeed/FSDP, vLLM/TGI/SGLang)Strong, reproducible research engineering: clean Python and disciplined experiment tracking that your collaborators can build onAbility to work independently on ambiguous and complex objectives, and to communicate clearly within a large project teamExperience generating training data and environments — synthetic data, agent trajectories/rollouts, and task simulatorsFamiliarity with inference-time scaling / test-time compute (search, self-consistency, verifier-guided decoding, long chain-of-thought)Experience with agent safety and guardrails: sandboxing, abuse/jailbreak resistance, and reliability for autonomous systemsA knack for interpretability and failure analysis — diagnosing why a model or agent fails, not just that it doesNotable open-source contributions and excellent technical writingPassionate about cybersecurity, with a firm understanding of the problem space — or passionate about applying your machine-learning skillset to a new domain such as cybersecurity (a security background is a plus, not a requirement)An independent self-starter who likes to take ownership and seeks out new challenges, and is thirsty for knowledge — never hesitant to step outside your comfort zone to learn new technologies, algorithms, and conceptsBenefitsEligibility for bonusesEquity grantsA comprehensive benefits package that includes health insurance, 401k and paid time offMarket leader in compensation and equity awardsComprehensive physical and mental wellness programsCompetitive vacation and holidays for rechargePaid parental and adoption leavesProfessional development opportunities for all employees regardless of level or roleEmployee Networks, geographic neighborhood groups, and volunteer opportunities to build connectionsVibrant office culture with world class amenitiesGreat Place to Work Certified™ across the globeCompany OverviewCrowdStrike is a cybersecurity technology firm that provides cloud-delivered protection for cloud workloads, identity, and data. It was founded in 2011, and is headquartered in Sunnyvale, California, USA, with a workforce of 5001-10000 employees. Its website is http://www.crowdstrike.com.