About the role

Assistant Vice President - AI Runtime Security

The role is responsible for defining and operationalizing governance controls that protect AI
systems at runtime, ensuring that AI, GenAI, and Agentic AI solutions operate securely,
predictably, and in compliance with regulatory and client expectations throughout their
lifecycle. This role focuses on runtime risk domains—including model misuse, adversarial
behavior, prompt and agent injection, data leakage, inference abuse, drift, and
unauthorized access—while partnering closely with engineering, platform, cyber security,
developers, businesses and clients
• Strong runtime protection for AI systems in production
• Reduced exposure to AI misuse, data leakage, and agent abuse
• Clear alignment to NIST AI RMF, security-by-design, and regulatory runtime
expectations
• Defensible governance posture for clients, auditors, and regulators
Responsibilities
• Define and own enterprise AI governance controls focused on runtime security,
monitoring, and enforcement for GenAI, LLM, RAG, and Agentic AI systems in
production.
• Establish technical standards for runtime threat detection and prevention, covering
prompt injection, agent manipulation, inference abuse, data leakage, hallucination
exploitation, and unauthorized model access.
• Ensure AI runtime architectures incorporate guardrails, policy enforcement points,
and telemetry collection across APIs, orchestration layers, model gateways, and
inference pipelines.
• Oversee implementation of continuous monitoring and observability mechanisms,
including behavioral monitoring, data and concept drift detection, usage anomalies,
and output risk indicators.
• Institutionalize governance requirements for runtime risk response, including
alerting thresholds, automated containment, escalation workflows, and integration
with enterprise cyber and incident response processes.
• Partner with platform, security, and MLOps/LLMOps teams to embed runtime
controls into CI/CD pipelines, model deployment workflows, and API management
layers without impacting delivery velocity.
• Define governance expectations for secure AI operation at scale, including access
control, rate limiting, logging, explainability at runtime, and auditable control
evidence.
• Lead technical governance for high-risk and regulated AI deployments, ensuring
runtime behavior complies with internal policies, client contractual commitments,
and global regulatory expectations (e.g., NIST AI RMF).
• Act as the technical advisor for AI runtime risk decisions, advising executive
stakeholders and clients on production readiness, risk acceptance, and control
effectiveness.
Qualifications
• Bachelor’s or Master’s degree in Computer Science, Information/Cyber Security,
AI/ML, Data Science, or related field
• 10-15+ years overall experience, including 3+ years in AI governance, AI runtime
threat vectors and AI observability, monitoring, and drift management
• Proven ability to design and govern runtime guardrails using AI governance and risk
platforms (e.g., Credo.ai for AI inventory, policy enforcement, and risk assessment)
• Strong hands-on understanding of runtime monitoring and observability for AI
systems, leveraging LLMOps/MLOps platforms such as MLflow, Weights & Biases,
Datadog, Azure Monitor, CloudWatch, or equivalent, to track usage patterns,
behavioral anomalies, and model drift in production.
• Command of drift detection and AI behavior monitoring, including data drift,
concept drift, and output instability, using observability and model monitoring tools
(e.g., Aporia, Datatron, custom telemetry built on OpenTelemetry).
• Ability to operationalize AI runtime governance controls within CI/CD and
deployment pipelines, embedding security checks and enforcement into
MLOps/LLMOps workflows orchestrated through platforms such as Kubeflow,
Airflow, GitHub Actions, Azure DevOps, or Jenkins.