AVP, AI/ML Ops

Lincoln Financial•Radnor, PA

3d•Hybrid

About The Position

We are creating a new AI/ML Ops pillar to own the reliability, safety, and cost-effectiveness of every model, pipeline, and agent we run in production. This is fundamentally an operations leadership role: someone who has built monitoring, escalation, and change management discipline from the ground up during a technology or business transformation, and who can apply that same rigor to AI/ML systems. This leader will build and run the function that closes the gap between "built" and "running well at scale." They will report to the Chief Data & AI Engineering Officer and sit alongside the leaders of AI/Data Infrastructure, Data Strategy & Governance, Data & Delivery Platform, AI/ML/Agentic Engineering, and AI Native Intelligence Platform.

Requirements

Proven track record leading operations through a major transformation: building the operating model, monitoring and observability foundation, and escalation discipline for a function that didn't have one before, ideally in technology, platform, or shared-services operations.
Deep experience in change management: standing up change advisory processes, risk tiering, release governance, and rollback authority in a fast-moving technical environment.
Strong background in monitoring, observability, and notification and escalation design: knowing what to alert on, who gets paged, and how severity and ownership are defined.
Experience building incident response and postmortem practices, including on-call rotations and MTTD/MTTR reduction, in any production technology environment. AI/ML-specific experience is a strong plus, not a prerequisite.
Working knowledge of (or demonstrated ability to quickly learn) ML/LLM system failure modes: drift, degradation, hallucination, latency and cost blowups, and how to instrument against them.
Experience partnering with governance and risk functions to enforce policy in live systems.
Strong track record hiring, developing, and leading technical or operational teams, ideally from an early stage.
Comfortable operating with ambiguity and building process from scratch, not just running an existing playbook.

Responsibilities

Define the operating model for how AI/ML systems move from build to production: intake, change approval, release governance.
Build change management processes for model, pipeline, and agent releases, including change advisory review, risk tiering, rollback authority, and communication protocols.
Establish a clear ownership and accountability model (RACI) across Engineering, Ops, and Governance for every production AI/ML asset.
Build monitoring for performance, drift, latency, quality, and safety across production AI/ML systems.
Design the notification and alerting architecture: what gets flagged, to whom, at what severity, with what SLA for acknowledgment.
Implement in-production evals and guardrails; enforce governance policy at runtime in partnership with Data Strategy & Governance.
Define and own SLOs/SLAs for production models, pipelines, and agents.
Build on-call rotations, incident response, escalation tiers, and postmortem practice for AI/ML systems.
Own the detection and escalation path for misuse, abuse, or safety incidents.
Drive mean-time-to-detect and mean-time-to-recovery down as the estate scales.
Partner with Platform Engineering to stand up CI/CD, versioning, and rollback for models, pipelines, and agents, partnering with Engineering on execution.
Establish safe rollout patterns (canary, shadow, staged) in partnership with Engineering.
Own the retirement and deprecation process for models and agents no longer in use.
Own inference and compute cost visibility and optimization; build unit economics per model and agent.
Partner with AI/Data Infrastructure on capacity planning.
Define the operating model, team structure, and hiring plan for the pillar.
Build the feedback loop that routes production failure modes and cost signals back to Engineering.
Establish the pillar's KPIs and report on them to the Chief Data & AI Engineering Officer and broader leadership.