Tojvs6qotaw8cysmu27j

Cheng-Yang Jiang


Experience with natural language processing(NLP), data mining, machine learning, deep learning, big data, social media analysis, and web crawling. Proficient in Python and familiar with text processing.

Hopes to focus more on NLP, AI, data science and data engineer in future career.



Taichung, Taiwan
[email protected]

  :0983-181293

 Education

National Chung Hsing University, M.S., Computer Science, 2017 ~ 2019

GPA: 4.12/4.3

Ubiquitous Data & Intelligent Computing Lab.

Focused on the fields of natural language processing(NLP) and deep learning. 

University@2x

National Kaohsiung University of Applied Sciences, B.S., Electronic Engineering(Information group), 2013 ~ 2017

University@2x

 Skills


Programming

  • Python
  • C/C++
  • Java
  • Objective-C
  • SQL


Toolkits

  • PyTorch
  • Pandas
  • NumPy
  • scikit-learn
  • Matplotlib/Plotly

DevOps

  • Linux(Ubuntu)
  • Docker
  • Hadoop/Spark
  • Git

 Work Experience

Industrial Technology Research Institute, Jul 2018~Aug 2018

Intern

  • Design deep learning models using NLP skills to assist the company to participate in the "2019 Formosa Grand Challenge – Talk to AI" competition hosted by Ministry of Science and Technology(MOST).
  • Design web crawlers to extract data from Wikipedia and use it to train classifiers.
  • Improve the accuracy of speech recognition by 10%.

Company@2x

Steams System Inc., Jul 2016~Jun 2017

Intern

  • Develop iOS Apps in Objective-C.
  • Design a ride-hailing system.
  • Responsible for front-end UI and function design.

Company@2x

 Projects


  • Abstract Sentence Classification based on BERT 
    • Investigate the employment of the pre-trained BERT language model to tackle the sentence classification problem in the scientific abstracts.
    • Reformulate the sentence classification problem as a reading comprehension problem.
    • Outperform the existing best models by 4%-5% in terms of F-measure(F1) scores.
    • Yield state-of-the-art results in this field.
    • Skills: BERT、NLP、Deep Learning

Paragraph image 00 00@2x ebb59a6d9adb03673d06762584bb6a0cc401a7cc4bd081bb82ce6f841d95aa2b

  • Automatic Domain Classifier based on Wikipedia data

Projects 01 00@2x

    • Design web crawlers to extract data from Wikipedia.
    • Train a classifier that can distinguish domains.
    • Use natural language processing(NLP) skills to improve the accuracy of speech recognition
    • Intern project at Industrial Technology Research Institute.
    • Skills: Big Data、NLP、Deep Learning、Web Crawler、Classification

  • Ride-hailing system App

Projects 01 00@2x

    • Develop a ride-hailing system for iOS platform in Objective-C.
    • Responsible for front-end UI and function design.
    • Skills: iOS App development

  • Hadoop cluster computing system based on Docker
    • Set up a Hadoop environment on Docker, which can be used by more than 70 students at the same time.
    • Create open source teaching material, promote Hadoop knowledge, and teach students how to set up the entire system from scratch.
    • Customized Hadoop computing system to solve the workload of big data.
    • Skills: Docker、Hadoop

  • Taiwan Ptt word cloud
    • Design web crawlers to crawl about 700,000 articles from Ptt
    • Speed up with Spark.
    • Keyword extraction by TF-IDF.
    • Text visualization.
    • Public opinion analysis.
    • Skills: Big Data、Data Mining、Machine Learning、Web Crawler

Paragraph image 00 00@2x ebb59a6d9adb03673d06762584bb6a0cc401a7cc4bd081bb82ce6f841d95aa2b

  • The character relationship Network of Wuxia novel

    • Network graph for establishing relationships with characters in Jin Yong's wuxia novels.
    • Extract the correlation between characters through reverse index.
    • Network visualization.
    • Skills: Python、Text Mining、Visualization

Paragraph image 00 00@2x ebb59a6d9adb03673d06762584bb6a0cc401a7cc4bd081bb82ce6f841d95aa2b

  • Predict movie ratings based on user-based and item-based
    • The data set comes from the ratings of 4,000 movies by 6000 users, and there are about 1 million ratings.
    • Calculate similarity.
    • Movie ratings prediction.
    • Skills: Python、Machine Learning

  • Synthesize panoramic images based on affine transform and bilinear interpolation

    • Turn images of different angles and sizes to the same position.
    • Calculate new pixel values through bilinear interpolation.
    • Combine two panoramas of the same location, different angles and different sizes.
    • Skills: Python、Image Processing

 
Powered by CakeResumePowered by CakeResume