Over a million selling partners and 30,000 Amazonians make business decisions through the analytics foundation this Principal Engineer will own – across a $670B+ revenue ecosystem. This role is for the central data and analytics platform powering Amazon's selling partner experience, providing consistent, standardized access to thousands of business metrics across hundreds of slices and time periods through its curated data layer and analytics query engine, a metric governance catalog, and cross-domain analytics experiences serving both selling partners and internal Amazonians.

We are seeking a Principal Engineer to architect and lead the next generation of this platform – redesigning core infrastructure to support fundamentally new workloads including real-time agentic analytics, AI-native programmatic interfaces, and multi-tier query optimization at dramatically lower cost, all while maintaining zero-disruption continuity for mission-critical production systems that underpin 10+ applications with 99.99% availability requirements.

This is not greenfield work on a blank canvas, nor is it maintenance of a stable system. It is both simultaneously: rebuilding the engine while the plane is flying. The platform carries significant technical debt – a monolithic architecture with 300K+ lines of code accumulated over 6+ years on Redshift/Presto-based infrastructure – and is actively being modernized to serve an ecosystem that is rapidly shifting toward AI-assisted decision-making. The architectural choices made here will define how AI agents access and reason about Amazon's selling partner business data for years to come.

Three technical challenges define this role:

1) 10x the throughput at 1/10th the cost – on a live platform. You will redesign core query infrastructure to achieve 100K+ QPS while reducing cost-per-query to a 10th of current cost. The architecture must span three fundamentally different performance tiers – sub-100ms operational queries, sub-2.5s strategic planning queries, and bulk model-training workloads – each with its own optimization profile. The migration must be invisible to every downstream client.

2) Teaching AI agents to reason about business data, not just query it. Amazon is investing heavily in agentic analytics, and those agents need infrastructure that doesn't exist yet. You will design MCP-compliant endpoints, decision-scoped analytics agents, and knowledge bases that give AI systems semantic understanding of thousands of business metrics – bridging structured metric definitions with LLM consumption patterns. These are foundational choices with ecosystem-wide consequences.

3) One source of truth where there are currently many. Selling partners see conflicting metrics across Vendor Central, Seller Central, and Selling Partner APIs, which erodes trust and makes AI recommendations unreliable. You will drive consolidation under Ripple as the single governed platform, navigating competing priorities across multiple Amazon VP-level organizations while designing schemas flexible enough for cross-domain use cases – including joining order details with returns, inventory movements, advertising performance, and customer feedback at transaction grain – that no single team can solve alone.

This role requires a PE who operates across the full spectrum: writing critical-path code for foundational infrastructure, setting engineering standards through exemplary practice across a 40+ engineer team spanning four locations, and influencing senior technical stakeholders across various organizations to shape cross-Amazon architectural direction. Depth and breadth, in the same role, at consequential scale.


Key job responsibilities

• Architecture Strategy and Technical Vision: Own the end-to-end architecture strategy and three-year technical vision for the product portfolio spanning Data Services and Analytics. Ensure these systems evolve coherently – that Ripple's infrastructure serves analytics consumption patterns, that metric governance integrates seamlessly with the data model, and that the full stack works cohesively for both human decision-makers and AI agents.
• Platform Redesign and Hands-On Development: Architect and build V2 platform, designing and writing critical-path code for the new query engine, materialization framework, and multi-tier infrastructure (rapid-operational