Software engineer building core infrastructure for Amazon Ads' Spektr, an exabyte-scale advertising data lake. Day to day: designing and operating Java microservices on AWS-native infrastructure (Spark, EMR Serverless, Kinesis, Step Functions), building MCP-enabled AI agent interfaces, and automating CI/CD for petabyte-scale data pipelines.
Would you like to build highly available, scalable and distributed engineering systems for one of the largest data lakes in Amazon? Does Petabyte scale excite you?
Spektr is Amazon Ads' exabyte-scale central data lake: the unified catalog for petabytes of campaign, ad-serving, billing, and measurement data. Hundreds of teams and Amazon's advertiser-facing AI agents depend on it for discovery, ingestion, transformation, and querying at scale. If you want large-scale engineering with tangible business impact, this is the team.
We're rebuilding Spektr ground-up with an AI-first architecture, and building fully-automated domain agents that let customers manage their big data lifecycle in a single click. In parallel, we're adopting Apache Iceberg / S3 Tables with serverless compute, scaling horizontally via a data mesh pattern for 5× growth in volume and demand. If you thrive on greenfield experimentation and shipping things that matter to the business, keep reading.
As an SDE on this team, you will own the core platform services that power Amazon Ads data infrastructure, consolidating legacy systems into clean, API-first services that cut job creation time from weeks to less than a day. You will build MCP-enabled AI interfaces that let AI agents create and run data pipelines on Spektr at 99%+ success rate, directly shaping how GenAI discovers and consumes data across Amazon Ads. You will design and operate high-scale Java services on AWS-native infrastructure (Spark, EMR Serverless, Kinesis, Step Functions), processing petabytes of data daily at 99.9% availability. You will also drive deployment automation, replacing manual release cycles with CI/CD pipelines that accelerate delivery velocity for a platform serving Tier-1 advertising data systems.
Key job responsibilities
- Engage with key decision makers such as Product & Program Managers to understand customer requirements and brainstorm on solutions
- Design, code and deploy components and micro-services for the core job management pipeline
- Ensure testability, maintanability and low operational footprint for your code
- Participate in operational responsibilities with your team
- Innovate on AWS technology to improve latency, reduce cost and operations
A day in the life
- Focus on core engineering opportunities to guarantee system availability that matches our data growth
- Work with a skilled team of engineers, managers and decision makers to consistently meet customer demand
- Automate monitoring of data availability, quality and usability via simplified metrics and drive innovations to improve guarantees for our customers
About the team
The mission of the Ads Datalake team is to provide data that helps the advertising organization make informed analyses and decisions for our customers and to determine and deploy investments for future growth via a set of central and unified products and services.
Our team focuses on simplicity, usability, speed, compliance, cost efficiency and enabling high-velocity decision making so our customers can generate high quality insights faster.
We are a global team with presence across India and North America.