Artificial Intelligence Safety Institute, Oct/2024 - present ▪ AI Safety ResearchIoT Research Division, April/2017 - Sep/2024 ▪ Object-centric Representation Learning ▪ Video Prediction for Model-based Reinforcement Learning ▪ Multiagent Reinforcement LearningBioinformatics Research Section, Jan/2016 - March/2017 ▪ Machine Learning on EMR data to predict cardiovascular disease (CVD) risk

Rutgers Machine Learning Group, Representation Learning

Modeling of Airbnb Price and Demand▪ Implemented an automated script to collect about 200,000 Airbnb listings in six major cities over two months in Python.▪ Performed data analytic research to model potential sales of Airbnb listings using machine learning techniques, such as Regression, Neural Network, SVR, Random Forest in Python and R.▪ Clustering and classifying Airbnb listing based on context and sentiment analysis of user reviews in Spark and MLlib.Location Based… Show more Modeling of Airbnb Price and Demand▪ Implemented an automated script to collect about 200,000 Airbnb listings in six major cities over two months in Python.▪ Performed data analytic research to model potential sales of Airbnb listings using machine learning techniques, such as Regression, Neural Network, SVR, Random Forest in Python and R.▪ Clustering and classifying Airbnb listing based on context and sentiment analysis of user reviews in Spark and MLlib.Location Based Knowledge Sharing Platform by Crowdsourcing▪ Designed and implemented a mobile application and service in Java, MongoDB and AWS S3. Show less

Google TV Team, Google Korea▪ Merged 50+ code commits to the production branch after code review processes from at least one senior SDE. ▪ Reported 100+ bugs, fixed 40+ bugs and verified 30+ bug fixes using a bug tracking system. Mock Android Background Services for GoogleTV Search▪ Designed and implemented mock android background services for Test Driven Development of search features implementation in Java, Android NDK and Protocol Buffers.▪ Developed an internal tool… Show more Google TV Team, Google Korea▪ Merged 50+ code commits to the production branch after code review processes from at least one senior SDE. ▪ Reported 100+ bugs, fixed 40+ bugs and verified 30+ bug fixes using a bug tracking system. Mock Android Background Services for GoogleTV Search▪ Designed and implemented mock android background services for Test Driven Development of search features implementation in Java, Android NDK and Protocol Buffers.▪ Developed an internal tool to test the functionality of search features in Java and Android NDK.ResultViewer: Internal QA tool for automated test of GoogleTV OS▪ Initiated and built an internal QA monitoring tool to review test results for GoogleTV OS.Compatibility Test Suite for Google TV OS▪ Developed and released versions of test suite to test the standards of expected output and compatibility of GoogleTV OS in Java, Android NDK, Python and Bash. ▪ Refactored the system code in builder design pattern to improve maintainability, scalability and readability. Show less

Interactive Source-to-source Transformations for Flow-based Application▪ Reverse engineered and conducted profiling open source applications to identify component flow and bottlenecks using Understand from Scitools.▪ Investigated challenges of flow-based parallelism after manual parallelization attempts.

AGPS Team, Samsung Electronics Poland R&DAssisted Geolocation Positioning System▪ Conducted research on patents and academic papers about Assisted-GPS technology and developed a watchdog software of AGPS in C#.