Ad Endpoints – 8 billion monthly hits, daily peaks of 10,000 requests per second
+ Moved code from AWS Lambda to EKS yielding 90% cost reductions with monthly savings in the tens of thousands. Set up pod autoscaling and NGINX load balancers.
+ Tracked metrics and created dashboards to monitor endpoint traffic, data store writes, and cluster status and health.
+ Updated cookies to comply with Google’s third-party cookie changes.
+ Implemented unit, integration, stress, and baseline tests for completely untested code.
Ad Data Pipeline – processes 40 to 50 billion monthly rows
+ Supported a business-critical data pipeline that uses messaged based microservices with a focus on parallelization, idempotency, efficiency, and cost savings to prepare for increased data loads.
+ Designed a robust class architecture to process data of various formats with inconsistent rows. Onboarded many new sources of data. Implemented batched data store writes.
+ Handled various bugs and edge cases such as duplicate SQS messages and irrelevant files.
+ Implemented unit and integration tests for previously untested code.
