Jung-Bin Wang

Onsite software engineer with 4 years experience of developing and enhancing C++ system. Successfully develop fuzzy-logic-based algorithm for Taiwan radar data quality control. Solid background in mathematics and physical science. Graduate with distinct in physics. .

  Taipei, Taiwan  http://wangjb.github.io/         

Working Experience

2017.01 - 2021.04

Onsite Software Engineer  International Integrated Systems, Inc.

Onsite software engineer at Central Weather Bureau. Responsible for developing quantitative meteorological information algorithm and enhancing operational radar data processing system realized in C++. Cowork with meteorological colleagues, converting their FORTRAN code into C++ code for operational purpose. Using Python as script lang. Key projects:

  • Radar Data Quality Control based on Fuzzy Logic Algorithm on Taiwan Radar Network.
    • In this project, I develop a algorithm based on fuzzy logic to detect and remove unwanted non-meteorological signals, like interference pattern, arc artefact, ground and sea clutters, existed in Taiwan radar network. This work improves radar data applications like quantitative precipitation estimation. Two oral conference paper, and one abstract accepted in international conference.
  • Integrating wet radome effect algorithm(FORTRAN), bright band detection and correction algorithm into operational system.
  • Written an efficient C++ testing code for calculating statistics of radar data, to reveal location of clutters encountered in current radar scan strategy.
  • Fine-tuning quantitative precipitation forecasting algorithm, storm cell identification and tracking algorithm.

2014.04 - 2015.09

PhD Student  Leibniz Universität Hannover

Participates in MAIUS project, experienced building “Atom Chip” for realization of Bose-Einstein Condensate of K41 and Rb87. I learned:

  • Laboratory knowledge in clean room.
  • Complete PCB production procedure, including deposition, spin coating, exposure, developing, electroplating, etching, gluing, bonding, and also hetero-junction called mirror transfer which glues a slice of glass mirror onto the surface of electronic circuits.
  • Building electronic circuits to drive and read out output signal of the accelerometer.

2011.10 - 2014.03

Research Assistant  Institute of Atomic and Molecular Sciences, Academia Sinica

Join Ultracold Atomic Physics Laboratory conducted by Dr. Yu-Ju Lin. During this very early setup stage, my main tasks are:

  • Design feedback circuits for frequency-locking between 780 nm External Cavity Diode Lasers.
  • Setup optics for 1064 nm NdYAG high power laser.
  • Use EAGLE software to design and setup feedback circuits for supplying steady current.
  • Use LabVIEW and IGOR pro for implementing experiment and processing data.


National Cheng Kung University

MA, Physics

Physics major. Passed elective courses like “Introduction to Superconducting Physics and Quantum Computer”, Quantum Optics, Quantum Information.

Dissertation: "Non-abelian Geometric Phase and Quaternionic Hopf Fibration"

Supervisor: Prof. Chopin Soo

  • Honorary member of The Phi Tau Phi Scholastic Honor Society of The Republic of China (2011)

2009 - 2011

National Cheng Kung University

BS, Physics. GPA 3.8/4

Physics major. Have well theoretical training in classic physics and quantum physics.

Passed elective course like Programming(C lang, FORTRAN), Astronomy, Circuits, Electronics(Experiments), Special Theory of Relativity(1), General Theory of Relativity(1), Astronomy Physics, Plasma Physics, Statistical Mechanics. etc.

Passed compulsory courses in department of mathematics like Advanced Calculus(1)(2), Abstract Algebra(1).

Graduated with distinct in physics.

  • Outstanding Student for the academic achievement at 2008
  • Outstanding Student for the academic achievement at 2007
  • Outstanding Student for the academic achievement at 2006

2005 - 2009



Hi there, this is Jung-Bin Wang, was employed as a software engineer of International Integrated Systems stationed at Central Weather Bureau. I was responsible for developing meteorological information technique, enhancing and maintaining Taiwan radar data processing system.

During the second year of this job, I develop a fuzzy-logic based radar data quality control algorithm, constructed features and membership functions to identify and remove non-meteorological signals existed in Taiwan radar observation network, like interference pattern, ground clutter, sea clutter, etc. This work has two conference papers, also appear on 2020 ERAD conference.

My daily works include maintaining the backbone of Taiwan radar data processing system, which is realized in C++ language. From radar raw data decoding, radar quality control, radar data application (QPE, PID, etc.), to faster extracting radar archive data, and debugging, I have a full-stack working domain knowledge of radar data application, sophisticated in exploring features in space-domain and time-series data, experienced with algorithms like watershed segmentation and Kalman filter. On the other hand, I also write and rewrite FORTRAN programs to work with colleagues and use Python as scripting language for testing or data visualization. 

I am highly confident on my self-learning ability. In college years, I prompt myself to grasp the opportunity to build the front-end website for the department, which is still working well now. When as a research assistant in Academia Sinica, I build the phase-locking circuits and bias coil circuits, which are key components for providing laser frequency and bias magnetic field in high accuracy in order to cool Rubidium 87 atoms down to micro Kelvins to realize Bose-Einstein Condensate.

Thanks to the well academic training in mathematics and physics, I can tackle well software and hardware problems, for example, Fourier transform technique just plays a key role in removing interference signals in Taiwan radar data network. I am a curious person who eager to know more. Keep learning on Coursera, I got certificates on deep learning related courses. Also play in Leetcode to strengthen myself to produce high quality algorithm in work.

I got following Coursera certificates on theoretical and practical knowledge on machine learning deep learning.

  • TensorFlow: Data and Deployment 
    • I learn to deploy trained models to web-based environment using TensorFlow JS, device like android, iPhone, Raspberry Pi, and micro-controller like ARM Cortex using TensorFlow Lite.
    • I learn how to efficiently using TensorFlow Datasets architecture to setup data pipeline(ETL, extract, transform, loading) for routine model training.
    • I learn how to speed up training cycle with data cache and prefetch, and diagnosing bias/variance issue during training phase by using TensorBoard.
  • TensorFlow 2 for Deep Learning
    • In this fruitful specialization, I learn Probabilistic Deep Learning in TensorFlow Probability framework.
    • I learn how to implementing probabilistic distribution using bijectors, and transforming distributions through normalizing flow, and setup Bayesian Neural Network.
    • Accomplished a VAE(Variational Auto Encoder) capstone project by minimize KL divergence and reconstruction loss.
  • Deep Learning Specialization
    • I learn diverse models like DNN, CNN, LSTM, GRU etc. to model space domain or time series data.
    • I learn how to optimize hyperparameters like optimizers, batch size, etc, and understand bias/variance trade off during training.

Besides learning theoretical course from MOOC, I am also doing a side project to training key word spotting model deploy on ARM Cortex MCU. Through this project I understand practical constraints which AIOT will encounter on deploying applicable models on devices.

Powered by CakeResumePowered by CakeResume