Staff AI/ML Engineer

About the role

Dragos seeks a Staff Machine Learning Engineer to build production ML capabilities for ICS/OT cybersecurity, including threat detection, asset classification, behavioral analysis, anomaly detection, and NLP. The role spans cloud and resource-constrained on-premises deployments, real-time and batch pipelines, MLOps, observability, automated testing, CI/CD, Docker/Kubernetes, and adapting ML research to security. Qualifications require 6+ years of engineering experience (4+ in production ML), Python, SQL, and another language. LLM/RAG/NLP and cybersecurity/ICS/OT experience are relevant. Salary…

Requirements

• 6+ years of engineering experience, including at least 4 years implementing machine learning in production
• Strong Python and SQL plus one additional language such as Go, Rust, Java, or a JVM-family language
• Experience deploying ML systems with scikit-learn, PyTorch, TensorFlow, or Hugging Face
• MLOps experience with versioning, monitoring, orchestration, and high-reliability deployment
• Knowledge of data pipelines, stream processing, message queues, and medium-to-large datasets
• Knowledge of containerized deployment and cloud-native architecture
• Cybersecurity, threat detection, threat intelligence, or ICS/OT operations preferred
• Strong communication and cross-functional collaboration

What you'll do

• Design and implement production ML systems for cloud and resource-constrained on-premises environments
• Build and optimize models for ICS/OT threat detection, asset classification, behavioral analysis, anomaly detection, and NLP
• Develop robust ML workflows for real-time and batch processing
• Turn research prototypes into scalable production systems with OT detection experts
• Establish data contracts and ML pipeline observability, monitoring, versioning, and deployment practices
• Improve ML infrastructure with automated testing, CI/CD, and containerized deployment
• Adapt current ML research and open-source models to cybersecurity applications
• Optimize production model latency, accuracy, and resource use

Where this was posted

• LinkedIn

  last seen 2h ago
  View on LinkedIn