Staff Security Software Engineer, AI Security

About the role

At Databricks, we are at the cutting edge of AI and machine learning, and we are looking for a seasoned technical leader to enhance the security of our AI capabilities. In this pivotal role, you will be responsible for establishing and enforcing security protocols tailored for our AI systems. You will confront unique security challenges in a fast-paced and dynamic environment, ensuring that our AI technologies remain robust and secure.

Key facts

What you'll do

• Spearhead adversarial testing and red team operations on live AI systems, which includes foundation model APIs, agentic workflows, and the infrastructure used for model serving.
• Perform comprehensive security architecture assessments for intricate AI functionalities, pinpointing and addressing vulnerabilities related to prompt injection, data isolation, and model integrity.
• Create and implement automated tools and frameworks that facilitate scalable security testing and analysis for AI applications.
• Work closely with AI and machine learning engineering teams to integrate security best practices throughout the entire development process.
• Establish technical guidelines and standards for AI security across the organization, leading collaborative remediation initiatives.
• Provide mentorship to fellow engineers on offensive security methodologies and principles of secure AI design.
• Collaborate with cross-functional teams to ensure that security considerations are embedded in the design and deployment of AI solutions.
• Stay abreast of the latest trends and developments in AI security to continuously refine and enhance our security posture.

Requirements

• A minimum of 7 to 10 years of experience in offensive security, AI/ML security research, or product security engineering, with a proven track record of leading security efforts for complex systems.
• In-depth knowledge in at least two areas of AI security, such as security for large language models (LLMs) and generative AI (including prompt injection and jailbreaking), security for AI agents and orchestration, ML infrastructure and serving security, or governance and privacy of AI data.
• Demonstrated capability to design and execute adversarial attacks on operational AI systems.
• Strong grasp of AI/ML platform architectures, including aspects of model training, serving, and integration.
• Proficiency in Python for analyzing machine learning code and scripts, along with familiarity in at least one additional programming language like Go, Java, Scala, or Rust.
• Experience in creating automated security tools specifically for AI systems.
• A history of influencing product architecture and driving security enhancements across teams.
• Exceptional communication skills, with the ability to convey technical security risks into practical guidance for diverse stakeholders.
• A realistic approach to risk assessment, able to differentiate between exploitable vulnerabilities and theoretical issues.
• Expertise in at least one major cloud service provider (AWS, Azure, GCP) and its corresponding AI/ML security framework.

Nice to have

• Contributions to research or presentations at AI security conferences or relevant venues.
• Familiarity with AI security frameworks such as the OWASP Top 10 for LLMs or MITRE ATLAS.
• Experience with tools and platforms like MLflow, Unity Catalog, Delta Lake, or the internal workings of Databricks.
• Possession of an offensive security certification, such as OSCP.
• An academic or research background in machine learning, adversarial machine learning, or AI safety.

Skills & tools

• Python
• AI/ML Platform Architectures
• Cloud Security (AWS, Azure, GCP)
• Offensive Security Techniques
• Threat Modeling
• Development of Security Tools

Practical notes

• This position is available to candidates located anywhere in the United States.
• Databricks provides a comprehensive benefits package along with various perks to support employee well-being and professional growth.

Join us at Databricks and play a crucial role in shaping the future of AI security while working in a collaborative and innovative environment. Your expertise will help us safeguard our AI technologies and contribute to the responsible advancement of machine learning solutions.