Cybersecurity Data Scientist

Your growth matters to us - explore our career development opportunities.

Connect with others in our people-first culture and enhance our collective ingenuity.

Learn how we’ll support you as you pursue a balanced, fulfilling life.

Discover what to expect during your journey as a candidate with us.

As a Cybersecurity Data Scientist, you will operate as a hands‑on technical contributor and applied research leader responsible for designing, developing, and operationalizing data‑driven and AI‑enabled solutions for Booz Allen’s Cyber Operations teams. This role emphasizes execution and delivery, turning security telemetry, threat intelligence, and analyst workflows into production‑grade models, detections, and decision‑support capabilities that measurably improve prevention, detection, response, and recovery outcomes.

You will bridge data science and security operations by translating analyst needs, threat models, and incident learnings into reproducible data pipelines, feature sets, ML/LLM models, and evaluation frameworks deployed across cloud, network, endpoint, identity, and application telemetry domains. You will originate, facilitate, and lead cross‑functional efforts to mature AI‑enabled cybersecurity capabilities, including detection engineering augmentation, alert triage, threat hunting, and SOC automation, while guiding teams through threat‑informed model development, secure‑AI engineering, and responsible AI practices.

Perform model and solution reviews, provide technical direction for complex analytics initiatives, including SIEM, SOAR, and EDR data science integrations, cloud‑native security analytics, and GenAI tooling for analysts, and translate findings into actionable, measurable implementation plans. Leverage strong analytical, statistical, and communication skills to assess complex security and business problems, align technical and non‑technical stakeholders, and drive decisions to closure in support of Booz Allen Hamilton’s critical enterprise infrastructure, go‑to‑market platforms, and mission operations.

The ideal candidate for our Enterprise Cybersecurity team is technically inclined, intellectually curious, and adaptable, with a strong cyber‑defense mindset. They thrive in a fast‑paced, dynamic environment and are continuous learners who actively seek to understand complex challenges, ask thoughtful questions, and look beyond the obvious to identify innovative and effective ways of working. They bring a security‑first perspective, analytical problem‑solving skills, and the curiosity and aptitude to continuously evolve as threats, technologies, and mission needs change.

This position is located in McLean, VA.

What You’ll Work On:

• Design, build, and deploy custom AI/ML solutions for cybersecurity, including supervised and unsupervised detection models, anomaly and behavioral analytics, NLP on security text, retrieval‑augmented generation (RAG) pipelines, agentic workflows, and LLM‑assisted analyst tooling, and operationalize them end‑to‑end: data ingest, feature engineering, training/tuning, evaluation, deployment, monitoring, and retraining.

• Engineer scalable data pipelines over security telemetry, including logs, EDR, network, identity, cloud, and threat intel, to produce high‑quality, labeled, and feature‑rich datasets that power detection, triage, and hunting use cases.

• Apply rigorous experimentation, statistical analysis, and evaluation methods, including precision/recall, drift, calibration, A/B testing, and backtesting against historical incidents to validate model performance, reduce analyst burden, and quantify operational impact.

• Apply secure‑AI and MLSecOps engineering practices throughout the AI/ML lifecycle, including model and data protection, prompt and inference risk mitigation, evaluation against adversarial inputs, including evasion, poisoning, and prompt injection, and responsible AI controls.

• Integrate models and analytics into security tools and workflows, such as SIEM, SOAR, EDR, IAM, CSPM), extending detection logic, enrichment, and response playbooks with custom ML/LLM capabilities where…