AI Satety Ops

• Built a full-stack portal to help swiftly respond to and mitigate incidents by quick patching with Docker.

• Maintaining an inference service hosting thousands of ML safety models with Azure ML and Tensorflow.

• Implemented a feature to evaluate the impact of quick fix on ML models on Azure Kubernetes service.

• Onboard ImageGen and Sora to the Orchestrator service to provide logging and annotation for OpenAI new models

• Expedited the deployment of quick fixes from central Azure SQL DB to all regions with Azure “Kafka” and CDC.

FortiEDR (cybersecurity threat detection, protection and response)

• Built a full-stack web application for customers to manage licenses, billing files and cybersecurity environment.

• Developed the backend with Python Django, PostgreSQL and Docker and the frontend with React.js.

• Leveraged RabbitMQ and multi-thread Worker to handle heavy-load tasks like creating an environment and uploading big files, improving the service availability from 99.99% to 99.999% and deployment and processing time by 80%.

• Deployed our cybersecurity services and microservices on OpenStack/Google Kubernetes Engine with Shell Script.

• Streamlined CI/CD processes with Jenkins, boosting operational efficiency across the development lifecycle.

Alexa Entity Resolution Infrastructure

• Built an AI core engine and infrastructure of Entity searching (movie and music) on AWS ECS from Applications.

• Created resources on AWS CloudFormation and CloudWatch with CDK and TypeScript to monitor system performance, running services and improving security, ensuring 99.999% availability as a Tier-1 service.

• Increased the coverage rate of unit tests and Integration tests up to 80% with Mockito, TestNG and JUnit.

• Wrote shell scripts to install and clean up/set up AWS resources, improving development efficiency.

• Ran load tests to find the best Auto-Scaling rules for our Service, greatly improve the scalability.

• Created SNS service between our service and the Redis cluster and built its auto-scale rule.

• Created a Redis Cluster for caching the searching result and APIs to read from and write to the cluster.

Alexa Feedback

• Built the backend of the self-service website with AWS Lambda, facilitating customers to onboard their feedback.

• Built a customer-facing service on AWS EKS, collecting feedback from customers for the Alexa.

• Migrate our cloud infrastructure from Migrate AWS to Native AWS.

• Upgraded our old Redis Cluster to an updated version with CloudAuth, enhancing service security.

• Helped 10+ teams onboard their services to Alexa feedback, collecting feedback from the customers.

• Created a RESTful API talking to ranking service and employed a ranking mechanism to service reject rate.