Seunghun Cho Email & Phone Number

- Designed recommendation models for banner ads to increase CTR using Neural Collaborative Filtering.- Designed LSTM-based item recommendation models using purchase history data. Collected raw data using Kafka, preprocessed it in Scala. - Developed user clustering model based on users' app usage and purchase history using K-Means.- Developed game agents with DNNs. Preprocessed terabyte-scale play data into trainable datasets using Scala. - Built friend affinity models using… Show more - Designed recommendation models for banner ads to increase CTR using Neural Collaborative Filtering.- Designed LSTM-based item recommendation models using purchase history data. Collected raw data using Kafka, preprocessed it in Scala. - Developed user clustering model based on users' app usage and purchase history using K-Means.- Developed game agents with DNNs. Preprocessed terabyte-scale play data into trainable datasets using Scala. - Built friend affinity models using social graph (neo4j) and neural network embeddings. Show less

대한민국 경기도 성남

- Developed high performing A.I. player for imperfect information card games (Go Stop, Chinese Poker) using reinforcement learning (DQN) and tree search.- Developed automatic music pattern generator for rhythm games using LSTM and Autoencoders.

Seoul, Korea

Built multiple edutainment apps for children- Plant science app ‘NAMOO’ has been featured by Apple over 150 App Stores worldwide including U.S. and China.- NAMOO was promoted by Apple with the biggest banner in the App Store, and chosen to be introduced in their retail shops. - Awarded by prestigious American children’s media: Mom’s Choice Gold, Parent’s Choice Silver.http://crayonfriends.comhttp://namooapp.com

- Developed Windows client for a large-scale medical solution, PACS by MFC/C++- Developed SDK, PC client for intelligent surveillance system