Charmgil Hong Email & Phone Number

Pohang-Si, Kr

Pohang, North Gyeongsang, South Korea

* Teaching Machine Learning, Database Management Systems, Data Structures, C Programming, and Intro to Engineering Design.* Leader, Handong AI Lab (HAIL).

Pittsburgh, Pa, Usa

* Lead as a core contributor in a real-time clinical outlier detection project for ICU patient monitoring (PI: Dr. Milos Hauskrecht)* Conduct predictive data analysis and modeling at various levels of granularity with a large volume of time series data (45k distinct medical time series of 80k ICU patients), on the cloud computing environment that University of Pittsburgh Medical Center (UPMC) operates, to deliver a high-performing medical/clinical error monitoring system* Collaborate with domain experts to gain insights from clinical practices and validate findings from statistical data analysis; Work with clinician to evaluate the outlier detection (error detection) performance and reliability

Pittsburgh, Pa, Usa

* Primary focus: Multi-Label Classification that aims to learn and predict multi-dimensional output (multiple class labels or response variables); Conditional Outlier Detection that identifies unusual output patterns given observed context* Developed probabilistic graphical models and algorithms that are optimized to make multiple, simultaneous decisions* Implemented and operated ETL and machine learning pipelines that perform data preprocessing and selection, feature generation, model building (training and validation), and outlier alerting in an automated and scalable manner (using Java, Scala, and Python)

Pittsrbugh, Pa, Usa

* Taught both graduate- and undergraduate-level Artificial Intelligence and Machine Learning* Taught undergraduate-level Algorithm Implementation, Intermediate Java Programming, and Formal Methods in Computer Science

Princeton, Nj, Usa

* Research Topic: Sensor data analysis and multi-target regression for effective facility maintenance* Proposed a structured multi-dimensional regression framework to model the reliability and failure of mechanical devices; Demonstrated 10-30% improvement, compared with the previous multi-task learning approaches, in predicting future states of target systems* Implemented an R package for structured prediction* Filed an invention disclosure to Siemens

Palo Alo, Ca, Usa

* Research Topic: Cost-sensitive decision-theoretic classification/rule mining for automated identification of scraps from manufacturing processes* Our proposed classification method outperformed the existing rule-based approaches by 55%* Filed a US patent application* Implemented a Java software package that supports decision making in class-imbalance and cost-sensitive scenarios

Pohang, Gyeongsangbuk-Do, Korea

* Planned and proposed strategies and methods for Official Development Assistance (ODA) to the Korean Government* Organized and administered international development projects that were funded by the UNITWIN/UNESCO Chairs Programme, Korean Ministry of Education, Science and Technology, and National Research Foundation of Korea

Doctor Of Philosophy (Phd), Computer Science

Thesis: Multivariate Data Modeling and Its Applications to Conditional Outlier Detection Advisor: Dr. Milos Hauskrecht

Bachelor Of Science (Bs), Computer Science, Graduated With Distinction