Currently working on the Advertising API Infra team. I develop systems and software that orchestrate, authenticate, and authorize calls to all Amazon advertising APIs, serving >40,000 requests / second at <100ms latencies.

As a key engineer on a team of 3-4, I developed and launched a large-scale offline diagnostics system for Amazon Advertising, processing 55 million advertising campaigns daily to identify issues and optimize performance. This system reduced ad support contacts by 30%, directly enhancing customer satisfaction and operational efficiency. I led the implementation and testing phases, contributed to the design, and delivered an API with 99.9% availability and 250ms p50 latency, providing real-time diagnostics for agencies worldwide.

I tackled critical challenges such as resolving latency and product bugs, implementing a retry mechanism on SQS consumption to prevent event loss, and scaling ingestion to handle 60,000 records per second. I also enhanced processing capabilities by learning and utilizing Spark SQL, achieving a throughput of ~1,300 campaigns per second.

Additionally, I developed key components of a recommendation system that generated $5 million in annual recurring revenue. I mentored two junior engineers and one intern, led ops and scrum meetings, and helped drive the team's success in delivering high-performance solutions.

With expertise in AWS technologies like CDK, Lambda, and ECS, I build scalable, reliable systems that contribute to Amazon Advertising’s growth and innovation.

Building micro-services with serverless native AWS architectures for Amazon Advertising. Extensively worked with AWS Lambda, Cloudformation, CDK, ECS/Fargate, etc. to build API-first services in Java with high availability and low latency.

Owned the design, implementation and testing of a front-end experience in the Amazon Advertising console. Launched a single page react application in the ads console, that allows advertisers to troubleshoot ad campaigns. This reduced campaign troubleshooting time from several days, involving ad support, to minutes.

https://advertising.amazon.com/resources/whats-new/diagnostics-tool

Worked on object detection and tracking in the Autonomous Driving Perception R&D team.

Worked in the Interactions & Analytics Lab with Dr. Matthew Shreve, on semantic understanding of 3D objects. Developed a fast annotation tools for 3D objects, and a novel method to extract 3D model parts using Augmented Reality. Explored the use of Augmented Reality in conjunction with Computer Vision and Deep Learning pipelines. The work resulted in a patent: https://patents.google.com/patent/US10896317B2/en