Data Scientist
Resume:
Vincy Ding
Data Science Consultant
Vincy is a self-motivated and open-minded data scientist with advanced programming skills in Python, TensorFlow, MySQL, and Cloud Computing.
******@*******.***
Toronto, Canada
SKILLS
Python SQL
Machine Learning
Tensorflow Keras
AWS Research
Spark Tableau
Web Scraping
ETL Pipeline
EDUCATION
2019
Deep Learning
Specialization
Coursera
2014 – 2016
M.Eng in Civil
Engineering
University of Toronto
2010 – 2014
B.Eng (Hons) in
Architectural
Environment
Engineering
University of
Nottingham
WORK EXPERIENCE
08/2018 – Present
Data Science Consultant
Two Minus Technology
Toronto, Canada
Two Minus Technology is a company focusing on research of algorithmic trading solutions for foreign exchange. Developed a foreign exchange scalping model based on time series and technical indicators. Transformed candlestick data into hundreds of technical indicators and rolling window features. Tuned parameters of XGBoost-gbree for the forex model that can automatically trade EUR/USD which improved the win rate by 12%.
Deployed a trading robot on AWS Server, which can automatically decide whether to buy/sell EUR/USD every 10 min.
07/2016 – 07/2018
Capital Planning Data Coordinator
BGIS
Toronto, Canada
BGIS (a subsidiary of Brookfield Business Partners) is a leading property management company that manages 320+ million sqft of property globally.
Automated compilation and separation of excel spreadsheets, resizing and compression of building condition assessment photos using Python scripts, saving 320 person hours annually. Identified the deficiencies and evaluated the conditions of building envelope, components, and MEP (Mechanical/Electrical/Plumbing) systems to offer repair or replacement recommendations and cost estimations for capital budgeting plans. Analyzed 100,000+ rows of retail and corporate building assets data for over 2000 client buildings for one of Canada's top banks to effectively provide data insights for future capital spending using MySQL, Microsoft Excel and Tableau. Managed several engineering construction projects by preparing the AutoCAD architectural details drawings for construction, conducting site inspections, deficiency reporting, contract administration, and certifying progress draws for invoicing. Conducted energy modeling for LEED-certified retail branches of a top national bank to predict the energy performance for the engineering consulting team. MACHINE LEARNING PROJECTS
Santander Customer Transaction Prediction (02/2019) Built a deep neural networks model with regularization to identify which customers will make a specific transaction in the future for a bank in the UK - achieved an accuracy of 90%. Flower Recognition (02/2018)
Built a convolutional neural networks model with data augmentation using keras (tensorflow-gpu as the backend) - achieved an accuracy of 86% for training data and 79% for test data in predicting those 5 types of flowers.
Credit Card Fraud Detection (12/2017)
Preprocessed the imbalanced dataset using both under-sampling and oversampling (SMOTE techniques) methods.
Built various predictive models (support vector classifier, k nearest classifier, XGBOOST-gradient boosting tree, deep neural networks) to predict the 'fraud' and 'non-fraud' transaction on both under-sampled data and over-sampled data.
Achieved 96% precision (recall:93%) for the 'fraud' category and 91% precision (recall:95%) for the 'non- fraud' category through the deep neural networks model on under-sampled data.
Contact this candidate