Lead Cybersecurity Architect - AI/ML Specialty

As a Lead Architect (VP), you will provide independent oversight while partnering closely with engineering, product, and control stakeholders to embed resiliency-by-design into AI systems and delivery processes. You’ll help ensure that AI-enabled solutions—including modern GenAI/LLM workloads and MCP-based service integrations—meet enterprise security, resiliency, and regulatory expectations without slowing innovation. Job Responsibilities: Lead resiliency-by-design for AI platforms: Define, standardize, and drive patterns for high availability, disaster recovery, backup/restore, capacity planning, and operational readiness for AI/ML and GenAI services. Partner across the firm: Collaborate with Lines of Business and platform/service teams to influence architecture, controls, and operating models that support AI value realization and enablement at scale. Enable responsible scaling of GenAI: Support teams in productionizing LLM-powered capabilities (e.g., retrieval-augmented generation, agentic workflows, evaluation/monitoring, safety guardrails) with reliability, traceability, and appropriate controls. Establish and assess standards: Ensure AI solutions align to enterprise security, resiliency, and external regulatory expectations; identify gaps, define actionable remediation, and drive closure through accountable owners and timelines. Operationalize oversight: Review solution designs, validate runbooks and SLOs/SLAs, and guide teams through readiness checkpoints (testing, failover exercises, incident learnings, and continuous improvement). Drive pragmatic outcomes: Balance governance and speed by creating reusable guardrails, reference architectures, and enablement guidance that engineering teams can adopt consistently across environments. Required qualifications, capabilities, and skills: Significant experience designing, delivering, and operating enterprise-scale platforms across on-prem and cloud environments. Strong grounding in AI/ML engineering and the realities of production AI (model lifecycle, data dependencies, reliability, performance, monitoring, drift, and change management). Hands-on understanding of GenAI/LLM systems and modern patterns (e.g., RAG, agents, prompt/model evaluation, safety controls) and how these workloads behave in production. Demonstrated strength in resiliency, availability, recovery, and data protection controls—especially for distributed systems and cloud-native services. Experience leveraging cloud AI and platform services (e.g., AWS, Microsoft Azure, Google Cloud) and integrating managed services into governed enterprise environments. Ability to operate as a strategic partner and influencer, driving alignment across engineering, product, risk, and control stakeholders. Excellent communication skills, with the ability to translate between architecture, risk/control requirements, and engineering execution—producing clear decisions, trade-offs, and documentation. Preferred Qualifications: Experience translating business objectives and governance/risk requirements into pragmatic engineering standards, reference architectures, and “paved road” enablement patterns. Experience with Amazon Bedrock or equivalent enterprise model access platforms (e.g., Azure OpenAI, Vertex AI), including identity, access control, and governance considerations. Experience with GenAI mediation patterns (e.g., MCP or comparable governed tool/service routing), including auditability, policy enforcement, and controlled connectivity to internal/external services. Experience with Temporal or comparable workflow orchestration platforms for durable automation, long-running AI operations, and agent/workflow reliability. Experience with sandboxed execution, coding agents, secure agent runtimes, or developer workspaces for controlled code execution and experimentation. Experience with token/cost telemetry, usage attribution, chargeback/showback, or FinOps data pipelines for AI platform economics and capacity management. Experience supporting regulated, sensitive, or client-confidential workloads, including strong practices around data handling, privacy, and operational risk.