Delivery Lead

EPAM Systems•Orlando, FL

About The Position

The Delivery Lead owns one value stream, which may consist of one to three offshore development pods. The team delivers using an AI-native approach that is spec-driven, gated, and supported by the AI factory; maintaining that approach is part of the role. The Delivery Lead is accountable for planning and executing delivery, managing dependencies and risks, tracking delivery and AI productivity metrics, and improving the stream’s cost and throughput. The role also owns the stream’s contribution to the complexity-point baseline and helps move pods from time-and-materials delivery to committed output.

Requirements

Experience leading delivery of multi-team software programs with external vendors in a client-facing role.
Track record of delivering to a plan and meeting commercial commitments.
Hands-on experience with AI coding tools and first-hand experience using AI-native ways of working in a team.
Comfort working with delivery, cost, and productivity metrics.
Strong communication with senior client stakeholders and sound judgment about when to escalate.

Nice To Haves

Insurance or financial services background.

Responsibilities

Estimate each approved project with product owners, the lead developer, and architects.
Turn estimates into plans covering milestones, sprints, pod capacity, and staffing against the backlog, using AI Complexity Points as the capacity measure.
Map dependencies across teams, value streams, and vendors, assigning an owner and date to each.
Execute against the plan, track progress and throughput, and re-plan when scope, capacity, or dependencies change.
Raise risks and schedule slippage early, explaining the cause and the decision needed.
Report throughput in complexity points, cycle time, defect escape, and rework rate according to the account cadence.
Own the stream’s part of the complexity-point baseline and move pods from time-and-materials delivery to committed output.
Move each new pod through the AI factory ramp.
Track AI adoption and productivity by pod and by person.
Track AI costs by pod, person, and ticket alongside the cost of the pod itself.
Run a continuous improvement program for the stream, identifying sources of slow or expensive work such as rework, review queues, dependency waits, model and prompt choices, and runs that cost more than the ticket is worth.
Set targets for cost per complexity point and cost per unit of work, and reduce those costs over time.
Pass factory improvements on to the pods and measure their effect.
Keep savings visible and tie them to EPAM margin on the output-based model.