👋 Hi, my name is Maia

🏫 Masters student at the University of CA, Berkeley studying Information & Data Science

🧑‍💻 Come from a background in tech consulting, machine learning, and data science for tech, retail & consumer product industries

🤝🏼 Proven success leading teams, navigating ambiguous challenges, and establishing trust with diverse clients

💻 Looking to collaborate on data science projects centered on consumer journeys, privacy, and personalized entertainment experiences? Connect with me!

I developed a Convolutional Neural Network (CNN) to detect diabetic retinopathy (DR) from retina images captured under varying conditions. To improve generalization and mitigate overfitting, I applied advanced image transformation and data augmentation techniques such as rotation, scaling, and flipping. By optimizing the CNN architecture and enhancing the training dataset, I aimed to achieve high accuracy in DR detection, contributing to more effective automated diagnostic systems for diabetes-related eye diseases.

In this project, I developed a sentiment analysis model using deep learning and NLP to classify IMDB movie reviews as positive or negative. Trained on 50,000 reviews, the model outputs sentiment probabilities, enabling automated audience feedback analysis. This project showcases how machine learning can efficiently gauge sentiment at scale for applications in market research, brand analysis, and content recommendations.

The project focuses on Decision Tree and Linear Regression models, primarily using Scikit-learn and other Python ML libraries. Through feature selection, model tuning, and cross-validation, I developed predictive models capable of forecasting revenue with a 10% accuracy margin.

This project features two statistical models that investigate the relationship between Spotify’s track audio metrics and track popularity using a dataset of over 30,000 songs. Leveraging proprietary Spotify audio features such as danceability, energy, and instrumentalness, initial findings suggest a statistically significant, though modest, relationship, with danceability explaining only a small fraction of the variance in track popularity.

Smart ad bidding plays a crucial role in modern marketing and retail, using advanced data analysis techniques can help optimize advertising efforts in real-time. This project leverages insights from multi-armed bandit algorithms to design an adaptive auction model that enables multiple bidders and companies to dynamically maximize payoffs and ROI in competitive ad environments. The project is rooted in algorithmic design, data structures, and optimization and seeks to optimize the balance between exploration and exploitation in bidding strategies.

This analysis examines the relationship between political affiliation and voting difficulty, using a Chi-Square test to determine if party affiliation is significantly associated with voting challenges. The application of this research is to identify party-specific obstacles and inform targeted interventions to reduce voting barriers, enhance participation, and promote an equitable electoral system.

This project conducts a comprehensive analysis of the World Bank's World Development Indicators to address a pivotal research question: How is primary school enrollment associated with labor force participation and unemployment rates across low-, middle-, and high-income countries, and does this relationship vary by gender? The study seeks to uncover the broader socio-economic associations, offering valuable insights for policy development aimed at fostering gender-inclusive economic growth.

• Methodologies: Machine Learning, Deep Learning, Natural Language Processing (NLP), A/B Testing and Experimentation Design, Data Algorithms, Statistical Modeling, Predictive Analytics, ETL Processes
• Languages: Python (Tensorflow, Pandas, Numpy, Scikit-Learn, XGBoost, Scipy, Matplotlib), R (Dplyr, Tidyr, Caret, Ggplot2), SQL (for complex querying and data manipulation, HTML (for visualization integration / web-based tools)
• Tools: Dataiku, Snowflake, PowerBI, Tableau, Git, Amazon Web Services (AWS) Cloud Environments, MS Excel
• Data Engineering: SQL / NoSQL databases (Neo4j, MongoDB, Redis)

• PMP: Project Management Professional
• AI-900: Microsoft Azure AI Fundamentals
• PL-300: Microsoft Power BI Data Analyst Associate
• Deloitte Certified Core Data Scientist & Prompt Engineer