Dec 2019 – Oct 2021 · 1 yr 10 mos · South Korea

• 1) Worked on a service that provides personalized news articles/blogs to the users. Users can also post their own content.
• 2) Developed backend API using spring to provide CRUD functionality to Runestone Alphausers. Designed database schema for MySQL for backend API.
• 3) Surveyed and integrated NSFW filter to report inappropriate images. Optimized multimedia based API by modifying its design and implementing image pre-processing and storage in S3.
• 4) Structured and implemented Elasticsearch Indexes and Queries to provide content search feature to the users.
• 5) Built API Gateway using netflix’s Zuul in Spring, implementing features like OAuthprotocol, access token authorization and caching.
• 6) Modified Deep Interest Network using Python Keras to speed up the ML model training time. Improved the recommendation results by utilizing existing user features.

Sep 2018 – Dec 2019 · 1 yr 3 mos · South Korea

• 1) Analyzed URL and app usage data using Pyspark and deduced usage patterns based on day, location and time to provide a quick access list.
• 2) Developed unsupervised recommendation algorithm in Python for utterance & news using word2vec and user interests by mapping the item data to user interests.
• 3) Designed an algorithm to maintain across domain user vector for multiple recommendation fields using matrix transformation.

May 2017 – Jul 2017 · 2 mos · South Korea

• Implemented exponential family embedding paper to generate movie embeddings using user-movie data and proposed a recommendation algorithm. Received a job offer based on internship by Samsung Electronics.