Led and worked in several backend projects in the Risk & Fraud team, including latency improvement, reliability redesign, fraud pattern investigation and modeling, and mentorship.

Drive Faster! (Application)

An open-source application to show real-time driving ETA in the menu bar, hands-free.

• Created a Google Cloud Platform (GCP) application and set up Distance Matrix API.

• Developed a script with Python for API integration with rate limit in consideration.

• A real-time driving ETA is displayed in the menu bar with an adaptive strategy on time and date.

RiceBook (full stack)

a Facebook-like social network on Single Page Application w/ Restful API, MongoDB, etc.

• Designed the data model and UX with functionalities including registration, authentication, profile, newsfeed, post, etc.

• Constructed frontend components with AngularJS and styled with CSS and Bootstrap.

• Implemented backend with Node.js for request handling and MongoDB for data storage.

• Implemented test infrastructure with Karma (frontend) and Mocha (backend).

Minter (website + data index w/ retrieval)

A full-stack project to get daily tweets. An offline-version Twitter to serve my summer in China.

• Designed and developed a multi-facet Tweeter search engine with Scrapy.

• Used Apache Solr to index ~ 3M Tweets w/ multiple metrics (theme, keyword, region, etc.).

• Built an AngularJS website providing advanced and responsive search functionality.

The Owl Puzzle Game (Front-end)

An online challenging but amusing puzzle game.

• Designed the single-page application w/ navigation, sidebar and main section in styled CSS.

• Implemented randomization of image pieces with adjustable difficult using HTML and JavaScript.

The project started as a proof of concept for performance improvement in the Enterprise Risk & Security team. Building from scratch, the project certifies the performance gain by migrating the current in-house service to AWS. The team made the decision towards full migration based on the result.

• Designed and created a serverless workflow using AWS Step Function to detect transaction fraud.

• Implemented Util classes for JSON Parsing and HTTP GET & POST Request in the transaction process.

• Developed AWS Lambda Functions for in-house services integration.

• Initiated and drove the conversation with the downstream team for API endpoint compatibility.

• Implemented exception handling and failure recovery for system robustness in the state machine.

• Automated the workflow by using CloudWatch with dynamically scheduled tasks.