Senior Data Scientist, Applied ML

SpyCloud is on a mission to make the internet a safer place by disrupting the criminal underground. SpyCloud’s solutions thwart cyberattacks and protect more than 4 billion accounts worldwide. Cybersecurity is an exciting, evolving space, and being at the forefront of the fight to disrupt cybercrime makes SpyCloud a special place to work. If you’re driven to align your career with a fantastic mission, look no further!

We're looking for a Senior Data Scientist, Applied ML to design, build, and deploy models for critical cybersecurity use cases like incident detection and mitigation, fraud intelligence, and risk scoring.

You'll own the full model lifecycle — from data understanding and preparation through prototyping and deployment in production — and work closely with engineering, product, and research teams to turn complex problems into scalable, reliable systems. This role is ideal for someone who thrives in applied, hands-on environments where impact and collaboration matter, and who has genuinely owned data work end-to-end.

What You'll Do:

You will develop, train, and deploy models using real-world structured and unstructured data to power critical security features such as threat detection and alerting, entity resolution and risk scoring, and natural language-based tagging and classification. You'll build the preprocessing and feature engineering pipelines your own models depend on, and you'll own model monitoring and evaluation, designing feedback loops to continuously improve accuracy and effectiveness.

You'll be equally comfortable prototyping new approaches from scratch and taking existing prototypes — from our R&D team or your own experimentation — to production-grade reliability. This role sits deliberately at the intersection of research and deployment, not on one side of it: you'll take ownership of data validation, transformation, and pipeline health across the handoff points between research and production, not just within the boundaries of your own models.

Working closely with software and data engineers, you'll help productionize models in modern cloud-native environments like AWS.

This role is highly collaborative. You'll partner with product managers and domain experts to define success criteria, rapidly prototype MVPs to test new features or signals, and work with the data engineering team to access and understand diverse data sources, owning the transformation and validation steps throughout. Your input will also contribute to broader system design and architectural decisions.

Strong communication and documentation skills are essential. You will clearly articulate model design choices, tradeoffs, and outcomes to both technical and non-technical stakeholders, maintain thorough documentation for models, pipelines, and evaluation methodologies, and participate in model and compliance reviews and customer-facing discussions as needed.

Requirements:

• 4+ years of experience building and shipping models in production with direct, hands-on ownership of the data lifecycle around them
• Strong background in applied math (linear algebra, optimization, statistics) and machine learning
• Demonstrated experience leveraging Natural Language Processing (NLP) techniques for text classification, tagging, or entity extraction
• Proficiency in Python and key ML libraries: PyTorch, TensorFlow, scikit-learn, XGBoost
• Demonstrated experience building or maintaining data/feature pipelines (e.g., with Airflow, Spark, Pandas) as part of your own modeling work
• Comfort with model versioning and monitoring in production (e.g., MLflow, DVC)
• Working experience deploying models into cloud environments or containerized services
• Strong communication skills and the ability to translate complex problems into actionable solutions

Nice to Have:

• Deeper MLOps/DevOps/data engineering exposure: infra-as-code, CI/CD depth, etc.
• Familiarity with cybersecurity datasets or domains: threat intelligence, account takeover, ransomware, etc.
• Exposure to graph analytics, knowledge graphs, or cybersecurity frameworks like MITRE ATT&CK
• Background working with unstructured data (e.g., log files, threat reports, breach datasets)