Santa Monica, California, United States
Developed globally available distributed systems to protect the Disney Streaming suite from API fraud while maintaining a seamless experience for our users.
Built scalable Scala APIs to evaluate flagged user behavior and block fraudulent actors, processing over 500k requests per day.
Contributed to our resilient Apache Flink real time data streaming application that processes an average of 30 billion daily requests and analyzes incoming data for predictors of fraudulent behavior.
Lead a critical feature migration away from Kinesis stream based pipelines to Databricks jobs, ultimately cutting over $80k in monthly AWS costs.
Developed high-throughput infrastructure and applications to identify and respond to anomalous request patterns, record fraud decisions, and and persist bans across multiple levels of the Disney Streaming system.
Leveraged Redis (via the Lettuce Java client) for both fast access to metadata and critical ban decision making attributes, and separately as a Python Pub/Sub messaging system for forwarding ban rules to our NGINX reverse proxy to be used in blocking incoming requests.
Assisted in migrating the entire Fraud pipeline, including extensive coordination with teams across the engineering organization, to a 5th AWS region to support the global availability of the Disney Streaming suite.
Monitored and debugged production systems with distributed tracing, metrics, and Terraform based dashboards in Datadog.
Drove improvements for the Fraud team’s development lifecycle including expanding our test coverage to include end to end testing running in Jenkins, improving our CI/CD workflows for multi-region deployments, creating a Gatling load testing suite to refine scaling settings for new services, and writing response guidelines and escalation plants for our partners in tech ops to increase the speed and ease of responding to production alerts.