Tzu-Heng Huang

Software Engineer        Location: Taipei, TW          Email: [email protected]           : LinkedIn.Page

Background & Profile


Education

National Chengchi University          Sep. 2016 ~ Jun. 2020
  • Major in Computer Science (GPA 3.94 / 4.30)

  • Minor in Big Data Analysis and Fintech program

  • TA Experience: Data Base System and Data Mining

  • Research field: Data Mining (ML), Sensor Network, IoT, Time Series, and Edge Computing

Skills

  • Programming Languages: 
    Python
    , C++ / C, R, SQL, VBA, Shell Script, LaTex

  • DBMS: Postgres DB, MySQL, SQLite
  • Embedded Systems: Raspberry Pi (Raspbian), STM32 (Mbed OS), Linked 7688 (OpenWrt)
  • Other Skills: Tensorflow, Keras, ShinyApp, Flask Web, Git, PCB design (Eagle), LPWAN (Sigfox)
  • Languages: English, Chinese

Work Experience  (PR, authorized to work for any US or TW employer legally)

Jul. 2019 ~ Sep. 2019                      Argonne National Laboratory

Research Aide in Array of Things, @Chicago in the United States

Highlights: Data Science / Time Series Clustering / Regression Analysis

  • Array of Things (AoT) is an urban sensor network (more than 300 nodes have been deployed) in Chicago city, measuring environmental factors that impact livability in cities such as climate, air quality, pedestrian, floodwater, and noise (AoT Blog).
  • Developed ML approaches to calibrate temperature sensor data with the ambient light level.
  • Applied pattern identifications to improve the performance of calibration model by 25%.
  • Connected data API and transferred data to different sensor network platforms.

Feb. 2018 ~ Jul. 2020                              Academia Sinica

Research Assistant in Network Research Lab, @Taipei in Taiwan

Highlights: Edge Computing / Sensor Networks / Time Series Prediction

  • Contributed to the Location-Aware Sensing System (LASS), the biggest large-scale PM2.5 sensing IoT system in the world with more than 10,000 participating devices over more than 50 countries.
  • Implemented Seq2seq model for forecasting air quality on a real-time Edge Computing system.
  • Analyzed anomaly time series data and covered unexpected pollutant events automatically.
  • Developed a sensor hub to monitor air quality and recognized environmental sounds with CNN models.
  • Tested the first generation of MIT air sensors and enhanced data accuracy with calibration models.

Sep. 2019 ~ Present                              National Chengchi University

Research Assistant in the Data Mining Lab, @Taipei in Taiwan

Highlights: Data Mining / Machine Learning / Graph Theory

  • Explored the spatial-temporal correlation in sensor networks for a novel data estimation framework.
  • Analyzed the time series behavior in large-scale sensor networks with time-series clustering algorithm.
  • Resolved sensor inspection problems and reduced the cost of maintenance with graph theories.

Jul. 2017 ~ Apr. 2020                             National Chengchi University

Research Assistant in the College of Commerce, @Taipei in Taiwan

Highlights: Data Science / Employee Turnover / Data Visualization

  • Predicted employee turnover and organized an employee database based on end-user information.
  • Analyzed and visualized the working environment of short-term dispatch workers in Pegatron Factory.
  • Developed an online tool for Monte Carlo Simulation with ShinyApp frameworks.

Research Projects

Forecasting Air Quality with Seq2seq Model on Edge Computing, @Academia Sinica

  • Developed a sliding window-based time series forecasting model in Tensorflow Framework
  • Reduced burden on the cloud server and improved the efficiency and accuracy of prediction model.
  • Implemented a real-time forecast model on Raspberry Pi for Edge Computing.

Spatial-Temporal Correlation Modeling of Community-Based Sensor Network, @Data Mining Lab

  • Estimated air quality and discovered sensor spatial-temporal correlation among sensors. 
  • Reconstructed the missing data with correlation modeling in a large-scale sensor network.
  • Employed pattern recognition and time series clustering to improve the performance by 15%.

Calibrating PM2.5 Low-Cost Sensors in a Large-Scale IoT Environmental Monitoring Systems, @Academia Sinica

  • Established a regression-based calibration model and released an open-source API for citizens. 
  • Utilized professional weather stations as reference to correct low-cost sensor data. 
  • The proposal of project was accepted with Student Research Scholarship and approved by MOST.

Pattern Identification-Based Hybrid Calibration Model for Sensor Data, @Argonne National Lab

  • Proposed a data science approach to reduce radiative error and correct temperature data. 
  • Applied subsequence clustering algorithms with meteorological sensor data.
  • Developed ensemble learning techniques to improve proposed method and reduced the error by 25%.

Detecting Meaningful Events in a Large Scale Air Quality Sensor Network, @Academia Sinica

  • Developed correlation analysis from the field of spatial outlier detection and time-series data mining. 
  • Generalized an enhanced anomaly detection framework and identified pollutant events.

Key Sensor Discovery for Quality Audit of Air Sensor Networks, @Data Mining Lab

  • Optimized sensor quality auditing problem and reduced the wasting of manual cross-validations.
  • Revised K-center and Submodular algorithms on sensor networks for key sensor discovery.
  • This paper has been accepted by MobiSys '20 at Toronto, Ontario, Canada.

Software Projects

Dec. 2018 ~ Apr. 2019                                       PiM25 Community

  • Combined PM2.5 sensor and Raspberry Pi to develop a truly end-to-end IoT ecosystem.
  • Built a customer data visualization, outlook design, and an efficiently OTA update mechanism. 
  • The first Taiwan project published on the Magpi (an official Raspberry Pi magazine) and got accepted in HKoscon (2019) and COSCUP (2019) to give a short sharing talk.

Feb. 2018 ~ Mar. 2019         The First Generation of Environmental Sensor

  • Tested and deployed more than 70 environment sensors in Taiwan and detected the anomaly devices.
  • Provided a regression-based model for PM2.5 sensors to calibrate data and improved data quality.

Apr. 2019 ~ Jun. 2019                         Graduate Admission Prediction

  • Built a ShinyApp user interface to predict graduate admissions and visualized data analysis.
  • This project was awarded first place in the NCCU Data Science Contest.

Jan. 2020 ~ present                         Environmental Sound Recognition

  • Recognized environmental sounds with CNN-based models on an Edge Computing Sensor System.

Apr. 2020 ~ Jun. 2020                                 Twitter RecSys Challenge

  • Predicted and analyzed the user behaviors including reply, retweet, comment, and like on Twitter.
  • Developed machine learning pipeline with LightGBM, XGB, and Graph Mining algorithms for modeling.

Invited Talks

"The 24th of the Raspberry Pi Meetup".
Jan. 2019,  
Attendance 100+
"Raspberry Pi Jam".
Mar. 2019,  
Attendance 70+
"Techbang Magazine PiM25 Sharing".
Mar. 2019,  
Attendance 60+
"LASS Conference International Session".
Jul. 2019,  
Attendance 100+
"Internship Sharing Session in University".
Sep. 2019,  
Attendance 100+
"IoT Tutorial For Beginners".
Dec. 2019,  
Attendance 50+

Scholarships

Jul 2019 ~ Sep. 2019                          International Internship Program

A Student participating in an international internship program approved by National Chengchi University.

Jul 2017 ~ Mar. 2020                         Undergraduate Student Research

A Student participating in a research program approved by Ministry of Science and Technology (MOST).
Powered by CakeResumePowered by CakeResume
Powered by CakeResumePowered by CakeResume