Software Engineer Intern (Recommendation System)

Facebook                                                  May 2019 - July 2019

​

• Worked on the Video Ranking team to provide video recommendations for Facebook Watch tab

• Developed an exploration-focused video recommendation source that helps users explore new interests on Watch tab

• Built data pipeline and trained a multi-class model that predicts users' interests with over 70% accuracy

• Launched experiments and iterated on parameters such that this recommendation source maintains neutral watch time while increasing the level of exploration by over 98%

 

​

Software Engineer Intern (Computer Vision)

On Center Software                               May 2018 - Aug 2018

​

• Worked on image classification and object detection problems while heavily utilizing Google Cloud Platform

• Trained 2 image classifiers for classifying architectural images. Utilized transfer learning by training extra layers on top of Inception v3 using TensorFlow. Deployed and scaled the classifiers as a service using Google App Engine, Cloud Storage, Cloud ML Engine. Integrated the classifiers into a file control system

• Developed an auto-scoring tool to evaluate an object detection model. Doubled model evaluation productivity while also providing visualization. Deployed this tool as a web app using Python Flask, Docker, and App Engine Flexible

 

 

iOS Software Developer (External)

​Halliburton                                               Jan 2018 - May 2018

​

• Developed a location-based iBeacon iOS app to detect and report entrance and exits of employees wearing beacons within range of equipment, and to provide customized training/safety guidelines

• Connected the app with Halliburton’s backend database using Experience API (xAPI) for equipment usage tracking

​

​

Student Analyst

​Rice University Basketball                        Sep 2018 - Present

​

• Write scouting reports on upcoming opponents by watching game clips and analyzing data on Synergy to assist coaches on game plans

• Recorded advanced data such as contested shots, deflections, hustle points, and plus/minus both during games and during practice