University of Texas, Austin, TX

M.S., Computer Science

• Natural Language Processing
• Machine Learning
• Linear Algebra
• Distributed Systems
• Parallel Algorithms

College of Engineering, Guindy, Anna University, India

B.E., Computer Science

Research Scientist, Machine Zone Inc., CA, USA

• Deep Learning Object Detection: Used Faster RCNN Model in Tensorflow for object detection for extracting attributes in Image Assets. Leveraged ScaleAPI to annotate images for training custom object labels (Patent 7).
• Google TextAds Ranking: ML based scoring model for Google Text ads to decide whether to pause underperforming ads (Patent 6).
• Facebook Retargeting: Launch/manage Facebook Retargeting Campaigns using FBAds API.
• TV Ads attribution: TV Ads Attribution model to attribute installs to TV spots and measuring Install lift from TV Marketing Campaigns for Game of War, Mobile Strike games(Patent 4).
• Image/Video Feature Recommendation: ML based Feature Recommendation for Video Ads to boost performance for Mobile Strike (Patent 5).

ML/NLP Engineer, Machine Zone Inc., CA, USA

• Treding Chat topics: Analyze Trending Topics in game chats using Scalding.
• Search Engine: Built a search engine for game chat messages from Kafka using ElasticSearch.
• Correct for Rewards: Built in-game mechanical turk to make users correct errors in machine translations (Patent 1).
• Emoji Suggestion: Suggest emoji in real time as the user types in some message (Patent 3).
• Translation Correction: Built NLP system to automatically validate whether two messages in two different languages (10+ languages) mean the same with high accuracy (Patent 2).

Software Engineering Intern, Facebook Inc., CA, USA

• Developed GamerFriends API method through which a game could recommend friends with whom the user could join and play.
• Worked with Ticker stories for games in Facebook.