• Hello! I am

    Ke Xin

    Live as if you were to die tomorrow.

    Learn as if you were to live forever.

    - Mahatma Gandhi

  • Junior Data Scientist @

    Jewel Paymentech Pte Ltd Singapore

    Achievement

  • Projects & Online courses

    broken image

    Image Classification

    March 2018

    • 22nd in  Shopee-IET Kaggle Competition.
    • Applied transfer learning - Inception V3 on 38k images, trained a classifier to classify new 16k images into 18 classes using GPU with accuracy 78.9%.
    broken image

    Telegram chatbot - Mooboogle

    March 2017

    Deployed a bot program (Moboogle) on Telegram that helps user to navigate the price of an electronic product brand in http://whymobile.com/.

    broken image

    Data science specialization course

    2017

    Completed all 10 courses for data sciences specialization on Coursera in R Language.

    broken image

    Blockchain Basics

    2017

    Introductory-level course to learn about what blockchain is and what it might mean to each individual.

  • Educations

    broken image

    School of Physical and Mathematical Sciences

    Bachelor of Science(Hons)

    Mathematical Sciences, specialization in Applied Mathematics

    Minor in Computing and Data Analysis

    (Aug 2014- May 2018)

  • My Internship Experience

  • Companies I have interned at included

    broken image

    The China Navigation Company

    Cyber Security Intern

    • Pioneered Cyber Security Incident Response Plan for the company.
    • Implemented cyber security infrastructure framework.
    • Supports IT risk & security function. 
    • Review existing systems & put up recommendation for improvement. 
    • Analyse the readiness & incident reporting information and put into communication plan.
    broken image

    Linde Gas Asia

    Software Engineer / Data Analyst Intern

    • Deployed a raspberry pi to automatically snap the picture of gauge with other digitization team members.
    • Reduced procurement cost by predicting weekly shipment quantities using multinomial logistic classification, achieved an accuracy of 60%.
    • Utilized Tableau to visualize historical data and suggested a new way for procurement department in ordering deliveries that could save a significant cost per year.
  • broken image
    broken image
    broken image