Director, Security Products: Key Management

DigitalOcean is a cloud technology company seeking a senior security leader to secure AI systems and operationalize machine-learning-powered security at cloud scale. The role partners with security, engineering, and executive leadership to define AI security controls, lead adversarial testing and risk governance, and drive the company’s AI Security strategy and roadmap.

Responsibilities:

• Partner with Product Security and Engineering leadership to define how AI security controls are integrated into DigitalOcean’s customer-facing products
• Partner with Security Data Science and Security Engineering to build, deploy, and serve in-house AI/ML models, guardrails, tooling, and automations. Participate in ongoing monitoring of deployed models to detect drift, evaluate performance, and plan and implement improvements alongside model stakeholders
• Lead adversarial testing and AI red teaming, including jailbreaking, prompt-injection, evasion, and poisoning testing
• Lead security reviews of major AI product initiatives from design through launch, providing architectural sign-off and escalation guidance for high-risk decisions
• Threat modeling across the AI lifecycle: data ingestion, training/fine-tuning, evaluation, model serving, RAG, agent frameworks, and post-deployment monitoring
• Own and drive DigitalOcean’s AI Security strategy and multi-year roadmap, defining where we invest, what risks we prioritize, and how AI security evolves as the threat landscape and product portfolio change
• Architect and lead the AI Risk & Governance framework: establishing policies, standards, and controls for the responsible development, deployment, and monitoring of AI/ML systems across DigitalOcean and Cloudways
• Represent AI security at executive and board-level briefings; communicate risk posture, program progress, and strategic decisions to the CISO and senior leadership in clear, business-contextualised terms
• Define DigitalOcean’s position on emerging AI security topics including agentic AI, model supply chain risk, synthetic data integrity, and AI-enabled social engineering

Requirements:

• 10+ years of experience in cybersecurity, with at least 4–5 years focused on AI/ML security, adversarial machine learning, or security data science in a production cloud or technology environment
• Demonstrated track record of defining and driving AI or security programs at an organizational level — not just executing within them. You have owned a strategy, not just implemented one
• Deep, practitioner-level knowledge of adversarial ML attack and defence: evasion, poisoning, model extraction, membership inference, and prompt injection at both conceptual and implementation depth
• Ability to read, evaluate, and synthesise academic AI security research and translate it into concrete engineering guidance and organizational policy
• Experience engaging executive and senior leadership audiences on technical risk: you can translate complex AI threat scenarios into business-level impact and justify investment decisions accordingly
• Strong programming skills in Python and Go with a track record of building security tooling and automation, not just reviewing others' code. You can stand up adversarial testing, detection, and guardrail capabilities yourself
• Proficiency reviewing and critiquing AI/ML system architecture (data ingestion, feature engineering, training pipelines, model serving, continuous monitoring) made by ML engineers and data scientists
• Hands-on experience with AI red-team and defensive tooling and with model supply-chain frameworks
• Bachelor's or Master's degree in Computer Science, Information Security, Mathematics, or a related field — or equivalent depth of practical experience
• Fluency in OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF, ISO/IEC 42001, and the EU AI Act
• Prior experience in a cloud provider, security product company, or large-scale internet platform where AI/ML systems operate at significant customer-facing scale
• Familiarity with LLM security in production: securing RAG pipelines, agentic frameworks, and model APIs against prompt injection, jailbreaking, and data exfiltration