Salesforce

Software Engineer (3 months internship full-time)

Work at the infrastructure team that builds and operates a large-scale recommendation system, which supports over 700 businesses and over 2000 websites, using Python, Node.js, Cassandra, AWS, Terraform, Docker and Kubernetes.
Developed a tool that enables data scientists to test recommendation strategies directly with real-time recommendation service, by creating or updating containerized data science models onto AWS ECR; Implemented server-side to handle HTTP requests from realtime service to model containers.
Deployed model containers on AWS EKS cluster using Terraform, Kubernetes and Helm: Created deployment and service for model containers managed in pods; Configured routing to the corresponding service with ingress controller; Applied network policy to prevent unwanted traffic from model container pods into the database.
Created a new strategy to bundle EC2 target groups that speed up code deployment for realtime recommendation web service by more than 50%; Refactored existing Infrastructure as Code by migrating bash script to Terraform.