Description

AWS AI Services is one of the largest and fastest-growing business units within AWS, powering services like Amazon Bedrock, AgentCore, QuickSight, Q Business, Kendra, and Kiro. Our Data Engineering team builds the intelligence infrastructure that makes this portfolio measurable — from revenue attribution and launch telemetry to agent-generated business reviews that serve VP-level leadership weekly.

We are looking for an experienced, self-driven Data Engineer to join a team that operates at the intersection of data engineering and agentic AI. In this role, you won't just build pipelines — you'll design data platforms that power AI agents, build automated reporting systems that replace manual processes, and create the data foundations that prove business impact across a multi-billion dollar service portfolio.

You'll work with modern AWS-native data stacks (Glue, Redshift, Athena, QuickSight, Bedrock, SageMaker), build event-driven architectures with CDK, and contribute to agentic workflows that generate executive-level insights autonomously. You should be comfortable operating in ambiguity, designing data models from scratch for new services, and making architectural trade-off decisions that scale.

This is a high-visibility role. Your work will directly inform decisions made by VPs, GMs, and the CFO's office — from revenue unification mandates to enterprise deal velocity to AI adoption measurement.

Key job responsibilities
Design and build end-to-end data platforms for new AWS AI services — defining schemas, data models, ETL pipelines, and analytics infrastructure where none exists today
Build and maintain production ETL/ELT pipelines using AWS Glue, Airflow, Spark, and Python to source data from operational, commercial, and telemetry systems into unified data models
Develop agentic data workflows — automated reporting pipelines that leverage AI/ML to generate business insights, WBR summaries, and anomaly detection without manual intervention
Create event-driven data architectures using CDK, Lambda, SNS/SQS, and S3 event notifications to support real-time data ingestion and processing
Build executive dashboards and self-serve analytics using QuickSight that serve VP/GM-level leadership across multiple service lines
Own revenue data accuracy — implement and validate revenue attribution models, discount calculations, and financial data pipelines that feed CFO-mandated reporting
Design data models that support both operational analytics (feature adoption, customer health, churn signals) and financial reporting (revenue, billing, forecasting)
Collaborate with Product Managers, Finance, Service Engineering, GTM, and Data Science teams to translate business questions into scalable data solutions
Optimize pipeline performance — reduce runtimes, eliminate redundant processing, and improve SLA compliance across production workloads
Mentor engineers, contribute to team standards, and drive a culture of automation, code quality, and operational excellence

A day in the life
As a Data Engineer on this team, you will design data models for newly launched AWS AI services, build and deploy ETL pipelines to onboard telemetry and revenue data, and validate data accuracy across financial reporting systems. On any given day, you may be architecting a CDK-based event-driven pipeline, collaborating with Product Managers to define launch metrics, resolving data discrepancies surfaced by Finance, or optimizing production queries that feed into VP-level weekly business reviews. Your deliverables ship to production on a regular cadence and are consumed directly by senior leadership for strategic decision-making.

About the team
The AI Services Data Engineering team builds the data infrastructure behind AWS's Agentic AI portfolio — Amazon Bedrock, AgentCore, QuickSight, Q Business, Kendra, Kiro, and Transform. Our data powers the metrics and reporting that flow up to Amazon's CEO and CFO, supporting S-Team level visibility into Agentic AI revenue, adoption, and growth. We build automated WBR reporting with agent-generated summaries, revenue attribution models for multi-billion dollar pricing programs, and launch telemetry platforms for new GA services. We ship weekly, operate across multiple VP orgs, and value automation over manual work, clean data models over quick fixes, and engineers who own their domain end-to-end.