Nam Phung
Computer Science B.S., Statistics & Actualrial Science B.S. @ University of Minnesota, Duluth
Computer Science M.S. @ University of Minnesota, Duluth
Email: [phungle.thanhnam@gmail.com] [phung025@d.umn.edu]
Resume: [pdf]
Research interest: machine learning, data science, software engineer, evolutionary computation
Papers: [dblp] [Google Scholar]
Others: [CodeSignal] [HackerRank]
Teaching
University of Minnesota, Duluth
Office Hours:
- Monday @ HH314: 5pm - 7pm
- Wednesday @ HH314: 4pm - 5pm
- Friday @ HH314: 3pm - 4pm
Work Experience
Data Scientist @ BioTuring, Inc.
March 2020 - July 2022 @ HCMC, Vietnam
• Developer of BioTuring Browser (BBrowser) - A platform for analyzing and exploring Single-cell RNA sequencing (scRNA-Seq) data. The product is used by research groups from AstraZeneca, Sanofi, and Amgen to study human diseases, cancer treatments, drug development, etc.
• Developed and maintained application features for BBrowser including data exporting, visualization, search functions, and back-end services for data analysis. Increased monthly usage of scientific plotting feature by 50% in Q1 and Q2 of 2022.
• Implemented backend system that supports paralleling computationally intensive analytical tasks. The system is deployed on BBrowser and BioTuring Enterprise Server Platform of large pharmaceuticals from Japan, France, etc.
• Designed and developed an internal data pipeline that automate the process of releasing single-cell data to Bioturing public database.
• Assisted in maintaining the logging pipeline of BBrowser software using Elastic Stack.
• Researched and developed BBrowser cell type search algorithm. The algorithm takes an unknown input cell type population and search against a database of 80 million cells for populations similar to the input.
Machine Learning Engineer @ VCCorp
December 2019 - March 2020 @ HCMC, Vietnam
• Deployed GPU-accelerated deep learning video analysis models to production.
• Designed and developed data pipeline & back-end system to distribute uploaded videos from Lotus social networking platform into deep learning predictive models for data analysis.
• The distributed system analyzed approximately 80,000 videos on daily average. The analyses include video censorship, classification, copyright protection, text extraction, etc.
Undergraduate Research Assistant @ University of Minnesota Duluth
November 2018 - May 2019 @ Duluth, Minnesota
• Implemented Orion density-based outlier detection algorithm to identify anomalies within a dataset using Java, Watchmaker framework for evolutionary computation, and WEKA framework for clustering algorithm.
• Conducted performance analysis of Orion against several state-of-the-art distance-based outlier detection algorithms.
Undergraduate Research Assistant @ University of Minnesota Duluth
May 2015 - May 2019 @ Duluth, Minnesota
• Developed iOS mobile application for health care research to help monitor weight shifting exercising routines to assist patients with spinal cord injury to prevent pressure ulcers.
• Developed a threshold-based algorithm to detect user performing weight shifting exercises using pressure mapping system and Bluetooth 4.0 technology.
• Co-authored publications related to applied computer science in assistive technology used in health care industry.
Undergraduate Teaching Assistant @ University of Minnesota Duluth
August 2016 - December 2018 @ Duluth, Minnesota
• Organized teaching material for computer science lab section each week. The course teaches undergraduate freshman concepts in computer science using the Racket and C++ programming language.
• Prepared and grade written assignments, programming assignments, and exams.
Research Projects
Orion Outlier Detection Algorithm for Data Streams
Research project with Dr. Eleazar Leal that aims at analyzing and optimizing outlier detection algorithm for data stream. The target algorithm for the study is a density-based outlier detection algorithm called Orion from research papers at University of Oklahoma. Orion uses evolutionary computation and clustering algorithm to detect anomalies within a dataset. We conducted performance analysis of Orion against several distance-based outlier detection algorithms like Abstract-C, MCOD, and Approximate-Storm. We also developed and implemented new optimization solutions for Orion to reduce execution time in the evolutionary computation process.
Presentations
Phung, N. "Evaluation of the Orion Outlier Detection Algorithm for Data Streams", MinneWIC '19, Feb 2019, Duluth, MN. [link] [slides] [certificate]
Weight Shifting
Undergraduate research project with Dr. Arshia Khan to build a system consists of an iOS mobile application, a Raspberry PI, and a pressure mapping mattress that targets wheelchair-bound patients to help prevent pressure ulcers and study the correlation between the development of decubitus ulcers and sitting time. A Raspberry PI 3 computer connects the mattress and retrieves the data through a USB port using a Python script. The Raspberry PI acts as a peripheral device and transfer the data back to the iOS device using its built-in Bluetooth 4.0 module. The iOS app uses the CoreBluetooth framework to receive the data and uses it to visualize the real-time pressure mapping of users' buttocks area. The mobile app uses the data as input for a threshold-based algorithm to automatically detect weight shifting.
Papers
Khan, A., Reuter, M., Phung, N. "Wireless Solution to Prevent Decubitus Ulcers: Preventive Weight Shifting Guide, Monitor, and Tracker App for Wheel Chair Users with Spinal Cord Injuries (Phase II)", Proceedings of the IEEE Healthcom, Munich, Germany, Sept 2016. [link]
Khan, A. Phung, N. "Undergraduate Research in Assistive Technology: Design and development of a Preventive Weight Shifting App to Reduce the Risk of Pressure Ulcers in Wheelchair Bound Patients with Spinal Cord Injuries", ASME Journal of Medical Devices, 10(20):020927, Minneapolis, MN, May 2016; DOI 10.1115/1.4033277. [link]
Phung, N. "Assistive Technology: Mobile App Designed to Prevent Pressure Ulcers.", National Conference on Undergraduate Research (NCUR), Asheville, NC, 2016. [link]
Presentations
Phung, N. "Research in Assistive Technology: Design and development of a Preventive Weight Shifting App to Reduce the Risk of Pressure Ulcers in Wheelchair Bound Patients with Spinal Cord Injuries.", Design of Medical Devices Conference, May 2016, Minneapolis, MN.
Phung, N. "Undergraduate Research: Mobile App Designed to Prevent Pressure Ulcers.", UMD UROP Showcase, November 2015, Duluth, MN.
Medication Management for Patients after Open-heart Surgery
Undergraduate research with Dr. Arshia Khan from UMN Duluth to develop iOS mobile application that helps patients recently recovered from open-heart surgery to manage their medications.
Papers
Khan, A., Phung, N, Madden, J. "A Treatment Adherence Mobile Application Designed Specifically for Patients after Open-Heart Surgery", Proceedings of the International Conference on Health Informatics and Medical Systems, Las Vegas, NV, July 2017. [link]
Side Projects
Dots and Boxes AI with Reinforcement Learning
Machine learning project for senior computer science course CS 5751 - "Introduction to Machine Learning and Data Mining". Developed the environment to encode the game dots & boxes and implemented TD-0 Learning algorithm from scratch to build an agent that compete against a human player in a match game. [link]
Building Predictive Model for Weather Prediction in Szeged, Hungary
Statistics project that studies the relationship between temperature and other factors such as humidity, wind speed, pressure, visibility, date, etc by building predictive models using R language. Performed data pre-processing, data visualization, and classification analysis using two supervised learning algorithms: K-Nearest Neighbors and Random Forest. The classification rate of the temperature prediction in two models are approximately 72% and 80% respectively. [link]
Sudoku
Android sudoku game written in C# programming language with Unity game engine. [link]
Ecolibrium3 Solar Map Android Application
Developer in Android mobile application project that aims at raising the awareness of the importance of solar energy in Duluth area. The app provides the solar energy map of Duluth, MN to help improve the solar panel efficiency of the city. Implemented the solar map view controller using ArcGIS Sdk, designed database schema for the application using MongoDB, and implemented RESTful API for the server using Node.js. [link]
Problem Solver
Sophomore year class project from CS 2511: Software Analysis & Design. The program implements some graph search algorithms such as Depth First Search, Breadth First Search, and A* Search to solve some puzzles games such as 8-puzzle, blocks world, Hanoi tower, etc. [link]
Personal
When not working, I love to watch movies and read books. As a matter of fact, I'm a big fan of The Lord of The Rings (both the book and the movie). Besides that, my hobby also includes personal finance. I also like playing games (RuneScape, Lord of The Rings Online, etc) and watching soccer.