Senior ML Data Ops Manager, Prime Air

Amazon
Amazon

Software Engineering, Operations, Data Science

San José Province, Costa Rica

Posted on Jul 18, 2026

Description

We are seeking a Senior Data Operations Manager to lead the Data for Autonomous Aircraft (DA2) Labeling Operations team at Prime Air. In this role, you will own the end-to-end operational strategy, delivery execution, people leadership, and vendor management of a multi-workstream data annotation operation enabling algorithm development for safe, automated drone deliveries.
How do you get items to customers quickly, cost-effectively, and — most importantly — safely, in less than an hour? And how do you do it in a way that can scale? Our teams of hundreds of scientists, engineers, aerospace professionals, and futurists have been working hard to do just that! We are delivering to customers, and are excited for what's to come. Check out more information about Prime Air on the About Amazon blog.
If you are seeking an iterative environment where you can drive innovation, apply technologies to solve real world delivery challenges, and provide benefits to customers, Prime Air is the place for you. The DA2 program provides high-value, representative sensor data at scale through aerial collection and annotation. Within DA2, the Labeling Operations team based in San José, Costa Rica (SJO) executes data annotation across multiple workstreams including VTOL (Vertical Take-Off and Landing) semantic segmentation, Transit Phase Perception (TPP) including Forward Collision Warning (FCW), and Mid-Air Collision Alert (MACA) data labeling.
As a Senior Data Operations Manager, you will manage the labeling operation where the problem, opportunity, and strategy may not be fully defined — using your expertise and judgment to define metrics, goals, and long-term strategy. You will lead a team of managers and individual contributors, define team and cross-team goals by working backwards from customer problems, and establish the labeling operation as a center of operational excellence. You will be accountable for delivery, quality, efficiency, vendor performance, people development, and strategic planning — operating independently and influencing stakeholders across DA2 and Prime Air.


Key job responsibilities
Operational Strategy & Ambiguity

- Define metrics, goals, and long-term operational strategy for the labeling org in an ambiguous, undefined space
- Set team and cross-team goals by working backwards from customer problems
- Contribute to strategic planning (OP1/OP2) within the DA2 organization

Delivery & Execution
- Establish workflows and SLAs across all annotation/labeling workstreams (VTOL, FCW, MACA)
- Build and scale audit mechanisms that make outcomes auditable and quality measurable against goals
- Negotiate priorities, manage escalations, and mitigate long-term risks across teams
- Own leadership reporting (MBR, WBR, Kingpin Reviews) and flex resources to deliver results

Team & People Leadership
- Provide tactical and strategic management; may manage one or two managers plus ICs across multiple locations
- Build and adapt team structures to solve complex problems as priorities emerge
- Set performance expectations in 1:1s, lead calibrations, and run talent reviews

Vendor & Quality Management
- Define and enforce quality/productivity standards for vendor partners through SLAs and accountability mechanisms
- Own the Labeling team's Kingpin goals, driving to target via internal mechanisms and vendor enforcement

Cross-Functional Coordination & Influence
- Partner with stakeholders to define vision, prioritize projects, and communicate results through Amazon's writing mechanisms
- Serve as the integrator between labeling sub-teams and DA2/Prime Air teams (Collections, DCEPS Engineering, Perception Science, SWAN)
- Hold upstream teams accountable through documented SLAs and escalation mechanisms

Process Improvement & Operational Excellence
- Establish process improvement and operational excellence mechanisms, driving best practices across the team
- Proactively simplify labeling workflows and reduce exposure to classic failure modes before failures force change