Data Analysis and Machine Learning
Resume:
YuanJea Hew
Tucson, AZ **********@*****.*** 520-***-**** LinkedIn GitHub
EXPERIENCE
Graduate Research Assistant Tucson, Arizona
Center of Biomedical Informatics, University of Arizona August 2022 - May 2024
• Managed ETL processes for 10+ clinical trials and health studies, oversaw data analysis and management for over 100 participants, which improved data integrity and accessibility.
• Developed Python scripts to automate a data pipeline for ingesting CSV and JSON data into AWS tables, transforming and storing data as Parquet files in S3, and dynamically creating its tables in AWS Glue.
• Leveraged SQL queries through AWS Athena for streamlined data analysis on large wearable device datasets, enhancing data querying and processing efficiency.
• Created impactful data visualizations and interactive dashboards using PowerBI and Apache Superset to analyze performance trends, resulting in enhanced training strategies for the university football team.
• Presented an academic poster to stakeholders on the development and integration of medical applications for Fitbit, Apple Watch, and Google Fit devices on our data pipeline. Natural Language Processing Project: Sentiment Analysis Tucson, Arizona School of Information, University of Arizona August 2023 - December 2023
• Engineered a supervised neural network model for multi-label sentiment classification on over 30,000 social media text data entries using Python, Tensorflow and Keras.
• Mitigated model bias through calculation and assigning class weights in the training data, boosting underrepresented classes F1 score by up to 70%.
• Optimized hyperparameters by implementing Hyperband algorithm to enhance model performance, improving the overall F1 score by up to 4%.
Software Developer Tucson, Arizona
Biosenix LLC August 2020 - August 2022
• Developed fall detection algorithm on wrist-worn wearable product to enhance safety for independent elderly populations by classifying fall and non-fall events.
• Executed end-to-end machine learning pipeline, achieving ~ 90% accuracy in classifying wrist sensor data for fall detection using Python and Scikit-Learn.
• Deployed finite state machine algorithm and decision trees architecture using C++ into watch device.
• Utilized statistical methods to develop an activity detection feature on watch to classify user’s activity levels (sleep, sedentary, and active states) by using only accelerometer data.
• Showcased MLOps competencies by seamlessly integrating continuous Exploratory Data Analysis (EDA), feature engineering, model training, deployment, and monitoring throughout the machine learning lifecycle.
• Engaged in Agile software development sprints within a collaborative team environment in remote setting. Clustering Black Hole Images with Transfer Learning Tucson, Arizona Steward Observatory, University of Arizona February 2020 - December 2020
• Contributed to the comparison of black hole simulations by developing data analysis algorithms, facilitating the matching of simulated images with real black hole data.
• Leveraged a pretrained deep learning model in Keras for feature extraction on black hole images and employed K-means clustering to categorize over 100 synthetic images by shape and structure. EDUCATION
University of Arizona Tucson, Arizona
Master of Science in Data Science, Cumulative GPA: 3.8/4.0 May 2024 University of Arizona Tucson, Arizona
Bachelor of Science in Applied Mathematics and Astronomy, Minor in Physics May 2020 Relevant Coursework: Machine Learning, Artificial Intelligence, NLP, Cloud Computing, Data Mining, Data Visualization, Data Ethics, Linear Algebra, Vector Calculus SKILLS
Programming Languages: Python, R, SQL, C++, LaTeX
Frameworks/Packages: Tensorflow, Scikit-Learn, Keras, NumPy, SciPy, Pandas Cloud/Databases: AWS, MySQL Workbench, MongoDB
Data Visualization: PowerBI, Apache Superset, Plotly, Matplotlib, Visio Software/IDE: Git, Jupyter Notebook, VS Code, R Studio, Docker, Excel, MyDataHelps Languages: English, Mandarin, Malay
Contact this candidate