Data Python

Bertha Hsu

Canadian Citizen F* OPT Pittsburgh PA Mobile: +1-412-***-****

***************@*****.*** https://www.linkedin.com/in/bertha-hsu/ https://github.com/berthahsu0217 EDUCATION

Carnegie Mellon University, Pittsburgh, PA Aug. 2018 – May 2020 Master of Information Systems Management - with 30% scholarship Courses: Data Structures for Application Programmers, Object Oriented Programming in Java, Software Design for Data Scientists, Distributed Systems, Database Management, Intro to Machine Learning, Intro to Deep Learning, Intro to Artificial Intelligence, Natural Language Processing National Taiwan University, Taipei, Taiwan Sep. 2014 – June 2018 Bachelor of Arts in Economics

Courses: Micro/Macroeconomics, Advanced Econometrics, Financial Economics, Game Theory, Programming Design, Data Structures & Algorithms SPECIALIZED SKILLS

Programming: Python (Scikit-learn, PyTorch, NLTK, spaCy), Java, C++, R (Shiny), SQL

Tools: Git, AWS, Docker, Hadoop, Spark, NoSQL (MongoDB, Cassandra), Stata, Weka, Tableau, Qlik Sense

Specialization: econometrics, machine learning, deep learning, natural language processing, computer vision WORK EXPERIENCE

PAPER, (Winner of the EdTech Awards 2020), Montreal, Canada May 2019 – Aug. 2019 Data Scientist Intern

Reduced costs of overscheduling by developing a recommendation system of tutor units for daily shifts.

Achieved 96% prediction accuracy on online session demands with ARIMA, LSTM, Facebook Prophet models.

Exploited trends and seasonality in online user traffic using seasonal decomposition and time series techniques.

Identified efficient and inefficient tutors by writing a custom evaluation program to support the service team in monitoring tutor performance and improving essay editing service.

Wrote 100+ SQL queries and delivered 4 data analysis projects on platform user behaviours with interactive visualization. Routinely reported interesting insights and possible improvements to product managers. SYSTEX Corporation, SYSTEX Elite Internship Program, Taipei, Taiwan July 2017 – Aug. 2017 Business Intelligence Analyst Intern

Visually analysed 30+ company datasets and created BI dashboards for Data Center, using Tableau & Qlik Sense.

Trained in 10+ lectures/courses of latest technology trend, including Fintech, AI, IoT, blockchain, etc.

Participated in company visits and networking events in Microsoft Taiwan and Google Taiwan.

Contributed to market research and insight discovery in Financial System and Mobile Application industries, and delivered business proposals to managers of Information Service & Solution Center in weekly presentations.

Established 3 data analysis projects on Global Terror Attack, World Oil Trade, and Social Network with large datasets collected/scraped from online resources (i.e. Kaggle, Investing.com), and reported interesting insights. ACADEMIC & CO-OP PROJECTS

Automatic Jigsaw Puzzle Solver, CMU May 2020

Trained CNN-LSTM encoder-decoder model to reassemble fragments of N*N images with unknown position.

Improved accuracy to 88% on the CIFAR-100 dataset by applying data augmentation and Sinkhorn iterations. Best Moves using Deep Reinforcement Learning with Transformco (Sears), CMU May 2020

Implemented Deep Q-Learning models to find the best actions for the company in pursuit of profit maximization.

Developed a data simulator from ML models using years of transactions and offers that acts as the environment. Plastic Detection in Food Waste, CMU Apr. 2020

Located & classified plastic products in images of food waste, authorized by LOHAS Eco Environment Tech.

Optimized Faster RCNN model on custom data and achieving 85% of IoU between true and predicted boxes. Pittsburgh Crime Alert System, CMU Dec. 2019

Built a MapReduce application that did large computing of criminal offense data analysis in Hadoop platform.

Constructed an interactive interface that took entered crime info and displayed target locations in Google Map. Question Answering System on Wikipedia Articles, CMU Oct. 2019

Built Support Vector Machine Classifier to predict the answer type (Named Entity Recognition) of questions.

Improved accuracy to 86% by training GloVe algorithm on open data to generate word embedding vectors. Integrated Movie Ticket Booking Platform, CMU May 2018

Constructed a recommendation system using Singular Value Decomposition and Pearson Correlation models.

Scraped data from IMDb, local cinema websites with Python packages: BeautifulSoup and Selenium.

Produced sentiment analysis of tweets containing movie hashtags streamed from Twitter using RESTful API.

Contact this candidate