• Scaling Uber’s streaming data infrastructure using Apache Kafka to one of the largest Kafka deployments in the industry with trillions of messages per day

Columbia EventNet: http://eventnet.cs.columbia.edu/

• Developed a large scale structured concept library system with interactive ontology browsing, semantic event search and automatic detection of events and concepts embedded in user-uploaded videos

• Implemented the ontology data visualization using d3 library, and designed the back-end with searching function using Java Lucene and detection function using deep learning framework Caffe

• Built a data processing pipeline, transported data from data warehouse to AWS S3 and processed the data in AWS Elastic MapReduce using ETL manager, Java and Hive

• Utilized Hive on top of EMR and ran Hive script to build data families for generating insights from the data

2015 Global Intern Hackathon - 1st Place

• Developed an app for Amazon Echo to provide stock and market information automatically based on user requests

• Implemented functions on AWS Lambda in Node.js and retrieved information from Yahoo Finance in real time

Research on Super-Resolution Algorithm

• Designed a super-resolution algorithm which reconstructed HR image from several LR images by using the HR pixel as a parameter to fit a surface at each LR pixel and fused sampling values in the MAP fashion