A to Z Leave & Accom Tech, Amazonian Experience and Technology
Amazon
- Location
- US, WA, Bellevue
- Employment
- Full Time
- Work model
- On-Site
- Level
- Mid
- Posted
- 19h ago
Skills
About this role
We are looking for a Director of Software Development to lead the technology organization behind Amazon's leave and accommodations platform — the largest, most automated employer-operated absence management system in the world. This platform processes millions of leave and accommodation requests annually across 60+ countries, administers nearly $1B in annual benefit payments at greater than 99.99% accuracy, and serves over 1.5 million employees through their most consequential workplace moments. You will own the end-to-end technology strategy and delivery for two interconnected platforms that together manage the full spectrum of employee time-away and workplace modification. The core challenge: transforming a system built around human case managers into one where AI agents resolve straightforward cases autonomously, handle ambiguous cases with human oversight, and free human specialists to concentrate on the cases that genuinely require judgment and empathy. You will build autonomous adjudication engines, compliance governance agents, clinical job-matching systems, conversational AI experiences, and the composable service architecture that positions this platform as the foundation for a unified employee health experience — and eventually an external product. This is a role for a builder who thinks in systems, not features. Someone who can hold a 10-year architectural vision while shipping working software every 4–6 weeks. You will work at the intersection of AI/ML, enterprise platform engineering, regulatory compliance, and employee experience — at a scale and complexity no other employer or vendor has achieved. If you want to build something that has never existed before and that directly impacts millions of people during their most vulnerable moments, this is the job. Key job responsibilities AI-Native Platform Strategy: Own the technical vision for transforming leave and accommodations from human-processed workflows to AI-native orchestration — autonomous agents for straightforward cases, agentic workflows for ambiguous cases, human specialists for the rest. Autonomous Resolution: Lead development of an orchestration engine that transforms hundreds of leave policies into machine-executable workflows achieving majority auto-adjudication and zero-touch completion. Compliance & Governance: Own the structural separation of execution and governance — every case gets automated compliance audit regardless of whether a human or AI decided it. Composable Architecture: Drive API-first, tool-native design so all capabilities are callable, composable, and ready for broader platform integration and eventual external productization. Clinical Intelligence: Own AI systems connecting medical information to workplace decisions — job-matching, restriction extraction, dynamic documentation, and duration management. Conversational Experience: Drive conversational and voice-based leave management replacing callback queues with instant, contextual resolution. Scale & Modernization: Ensure durability at massive scale while completing migration from batch-processing legacy systems to event-driven architecture.
Team
Transformation: Lead a technology organization through the shift from volume-driven to complexity-driven staffing, building programs that upskill domain experts into AI Specialists. A day in the life You might review a PR/FAQ for autonomous adjudication, dive deep on return-to-work defect rates, partner with Legal on compliance implications of a new AI workflow, or work with AI Specialists building micro-automations. Your stakeholders span HR leadership, Operations, Legal, data science, and platform teams across Amazon's employee experience ecosystem. Your customers are over 1.5 million employees navigating health crises, new parenthood, and workplace modifications — plus the tens of thousands of HR professionals and managers supporting them. Problems range from badge lockouts on day one back to building the world's