AI Security Engineer

The Responsible AI Team at Databricks is seeking an AI Security Engineer to join their team in London, UK. This role involves performing security design reviews and red team engagements on new and existing AI models and systems, as well as conducting novel research in the AI security field.

Key responsibilities include:

• Conducting Red Team operations on live AI systems in development and production environments
• Investigating new and emerging threats to ML systems
• Creating tools, techniques, and processes for discovering and reporting vulnerabilities in AI systems
• Guiding secure model and system development through the SDLC
• Pioneering best practices for various facets of ML technology
• Collaborating with internal teams to advance operational security and monitoring procedures

The ideal candidate should have:

• Strong background in Machine Learning and Deep Learning with coding experience in libraries like TensorFlow, PyTorch, or SparkNLP
• Expertise in programming languages such as Python or C++ for coding and secure code reviews
• Knowledge of adversarial machine learning techniques
• Understanding of cybersecurity principles and tools for vulnerability discovery and exploitation
• Strong problem-solving skills and curiosity to develop novel attack methods against AI systems
• Excellent communication skills and ability to work in a team
• Typically 4+ years of experience or advanced degree (MS/PhD) with 3+ years of experience in the ML domain
• BS or higher in Computer Science or a related field

Databricks offers a comprehensive benefits package including private medical and dental insurance, pension plan, equity awards, enhanced parental leave, and various wellness and professional development benefits. The company is committed to fostering a diverse and inclusive culture where everyone can excel.

Join Databricks to contribute to the development of responsible AI practices and help shape the future of AI technologies that benefit society while upholding high standards of integrity, privacy, and social responsibility.