Shirley Wang
LinkedIn: https://www.linkedin.com/in/wxy829475/ 412-***-**** ada61k@r.postjobfree.com Education
New York University
Master of Science at Data Science 2017 - 2019 May
University of Pittsburgh
Bachelor of Science at Mathematics; Economics 2013 - 2017 SKILLS
• Machine Learning, Deep Learning, Natural Language Processing, Time Series Modeling, Statistical modeling, Data Analysis, Data Visualization, Database
• Python, SQL, Hadoop, PySpark, AWS, TensorFlow, PyTorch, MATLAB, Stata, Tableau, Plotly, Scikit-learn, Git WORK EXPERIENCE
Data Scientist, Genius Sports Jul 2019 - Current
Los Angeles, CA
• Apply Neural Networks/machine learning models on basketball game dynamic predictions for various betting markets
• Conduct advanced analysis and statistical tests on data, model performance, simulation and extrapolation via python codebase with AWS S3 buckets
• Cooperate with API engineering team and traders to deploy the model into production in October for NBA season 2019 Data Science Intern : Viacom Jan 2019 – April 2019 New York, NY
• Researched on state-of-the-art NLP models and transfer learning techniques: Elmo, ULMFit and Bert.
• Aimed to generate a semantic NLP pipeline, providing a series of outputs like context and emotion classification from text data to improve ads marketing
• Delivered competitive analysis of products on quantitative performance of social media platforms, extracted features and metric to understand audience reaction patterns
Data Science Intern : Apteo May 2018 – Aug 2018
New York, NY
• Developed prediction modeling for financial product: long-term stock price by working with a codebase in python
• Worked on infrastructure using AWS, EC2 and self-designed pipeline concentrated on NLP/deep learning models
• Conducted large-scale feature analysis, time-series data manipulation and network structure/metrics/loss function research based on existing neural network models, improved the MAE metric by 2.3% PROJECT EXPERIENCE
Neural Machine Translation (NLP) Fall 2018
• Built recurrent encoder-decoder model as baseline for two parallel corpus, Chinese-English and Vietnamese-English.
• Applied GRU, LSTM, Attention mechanisms in 2 versions, and replicated the Transformer model structure
• Improved the BLEU score 26%, 58% from baseline model for Ch-En and Vi-En task respectively NLP and Sentiment Analysis on Yelp Reviews (Machine Learning) Spring 2018
• Implemented machine learning algorithms Multinomial Naive Bayes model and FastText model to predict the Yelp star rating based on 2,000k text reviews
• Constructed deep learning Bi-directional LSTM with both word-embedding and sentence embedding (Skip-Thoughts)
• Utilized Local Interpretable Model-Agnostic Explanations (LIME) package to conduct sentiment analysis Automatic Image Colorization (Deep Learning) Spring 2018
• Preprocessed three classes of images e.g. food, natural and people and transformed them in gray-scale images.
• Adopted VGG-16 as the initial layer of the neural networks to extract information and trained a Generative Adversarial Networks (GAN)
• Tuned the hyperparameters and reached the 0.18 MAE for food images and 0.66 MAE for natural urban image