Xinyi Cao

Software Developer

***User Preference Service***

• Re-architected preference service in multiple partitions to adopt DynamoDB Global Tables. Used SQS to poll/process messages, keeping the preference in sync across all AWS regions globally.

• Properly migrated existing user preferences from S3 to DynamoDB.

• Enabled read/write features from/to Global Tables, which eliminated the use of a home-brewed data replicator in multiple regions.

***Global Navigation for AWS Consoles***

• Setup favorite bar and pin feature so that users can easily navigate around AWS services.

• Defined metrics (front-end, client side, and ops metrics) and monitors to track the performances. Established pattern and configured alarms.

• Used CloudWatch to build interactive dashboards for better data visualization and user behavior analysis.

• Support region and language modules to help release new AWS partitions.

***Console Customization Service***

• Designed the service that provides "Unified" settings which enables customers have consistent experiences across browsers and devices.

• Implemented Lambda handler to add A-Record and setup Route53 updating process for custom domains across accounts.

• Implemented infrastructure to create and validate ACM certs, and update to APIGateway.

• Automated IAM role creation in the Route53 account for efficiency and security purposes.

***TLS Migration***

• Replaced Netscaler-based hardware load balancers with an Amazon Network Load Balancer (NLB).

• Onboard to JLBRelay, managed deployments and provisioned load balancers.

SDE Intern

Strengthen the AWS Console-Navigation pipeline aim for CICD.

• Improved AWS Console Home integration tests with Selenium Webdriver. Increased the test coverage and reduced test duration from 50 min to 10 min.

• Added Canary tests for customized health checks and setup predictable metrics/alarms to continuously monitor login performance for AWS management console.

• Setup entire infrastructure (create AWS resources, config different partitions/endpoints, etc.) to avoid repeated manual managements.

Software Developer Intern

Weather forecast website

• Migrated partial feature to cloud from local servers to handle more requests from a growing market and reduce the cost.

• Collected weather data with Python Scrapy, feed raw data to a machine learning model for weather prediction in 7 days, implemented interactive visualization graph with d3.js and hosted the website for posting the result with AWS.

• Used AWS CloudFormation, CodePipeline, etc. to accomplish continuous integration and continuous delivery.

• Used AWS CloudTrail, AWS Config and GuardDuty to enhance security by avoiding the AWS configuration being changed maliciously.

• Applied MapReduce to analyze a large graph with Spark on Microsoft Azure.

Research Assistant

AirPnp: Parking Sharing System

• Developed an Android mobile app for searching and sharing available parking slot followed Airbnb model.

• Integrated Eventbrite API into application to recommend the nearby event for users.

• Used AWS DynamoDB to hold the real-time data and used API Gateway to generate the RESTful API.

• Added request throttling by using AWS SNS and SQS to send alert email to administrator.

Student Developer

Open-source Urban Energy Data Analysis

• Used Neo4j graph database tools to analyze nuclear and hydro power plant data.

• Built a self-contained web application on R Shiny that allowed users to process, explore, and visualize data.

Personal Healthcare Evaluation Application

• Developed an Android mobile app to collect user’s self-tracking data like heart rate, steps, sleeping quality, etc. from fit-bit.

• Synced data to AWS S3 and applied the recommendation algorithm to push daily healthcare suggestions.

• Added load balancer, Cloudfront and multi-availability zone to handle large number of requests globally.

• Used CodeCommit, CodeBuild, CodeDeploy and CodePipeline to accomplish the automation.

• Designed scatter plots and heat map for clients to visualize the healthcare report online.

EcoChoice Mobility: Routing Suggestion Based on Fuel Consumption

• Built the formula among different parameters that affect the vehicle fuel consumption.

• Scraped raw traffic / road condition data from websites with python scrapy and cleaned the data with OpenRefine.

• Adjusted Dijkstra’s algorithm with weighted graph to design the most fuel-efficient route.

• Built an interactive JS web application to visualize customized routes on a map.