Sophia Beebe

Brigham Young University of Idaho

GPA: 3.42

Teacher’s Assistant – Programming 310

January 2025 – Present

• Provide guidance and assistance to students in Programming 310, reinforcing concepts in various programming languages and software development methodologies.
• Advised 30+ students in C#, C++, and Git/GitHub workflows, reinforcing programming concepts, debugging strategies, analytical thinking, and software development practices.
• Collaborate with course instructors to review coursework, module assignments, and stretch challenges.
• Support students in building personal software portfolios by researching, experimenting, and publishing well-documented projects to public GitHub repositories.
• Utilize advanced computer science skills to help students solve complex programming problems and prepare for future internship opportunities.

Project Manager – Architectural Design

Elected Project Manager for CSE 430 (Software Architecture)

• Elected Project Manager for an Architectural design class of 25 people, to complete a multi-tier software architecture document for a Rideshare app.
• Coordinated a multi-tier development project, allocating resources, monitoring overall project health, and overseeing performance evaluations.
• Maintained daily communication with team leads to gather requirements, provide guidance, and ensure alignment with project milestones.
• Developed a comprehensive understanding of key system architectures (component diagrams, data flow diagrams, ect.) and how to apply them to meet specific software requirements.
• Practiced ethical tech leadership by understanding and learning all project aspects, while fostering team trust, and encouraging autonomy without micromanagement of a 200 page document.

Machine Learning XGBoost

Implemented and evaluated multiple machine learning models, including XGBoost and Random Forest, for price prediction. We conducted hyperparameter tuning to optimize model performance, and used RMSE as the primary evaluation metric to compare model accuracy.

• Data Preparation & Exploration: Cleaned and normalized large datasets, analyzed feature distributions (EDA), and managed class imbalance (about 9:1 “no” to “yes”).
• Results & Insights: Selected XGBoost over Random Forest due to a 30% lower RMSE ($93,118.98 vs. $133,739.94).
• Improved confidence in price predictions by reducing the average prediction error by $40,620.96.
• Achieved an R² of 0.89, demonstrating strong predictive capability.

Through this project, I gained substantial experience in machine learning workflows (data cleaning, feature engineering, model tuning), best practices in DMPR data privacy considerations, and the importance of balancing business needs with predictive accuracy.

Machine Learning 450 Project

Collaborated on an AI-driven telemarketing campaign analysis for Banco Federal de Financas to identify clients most likely to subscribe to a term deposit. Key aspects included:

• Data Preparation & Exploration: Cleaned and normalized large datasets, analyzed feature distributions (EDA), and managed class imbalance (about 9:1 “no” to “yes”).
• Model Development: Implemented a Random Forest Classifier (scikit-learn), tuned parameters (class weights, max depth) to address overfitting and emphasized true positives.
• Feature Engineering: Transformed categorical features and scaled numeric features for optimal model performance.
• Results & Insights: Achieved ~86% accuracy, revealing diminishing returns after multiple calls. Demonstrated the impact of economic indicators (employment rates, interest rates) on client responsiveness.
• Limitations & Next Steps: Recommended cross-validation, deeper hyperparameter tuning, and continuous monitoring to improve real-world performance.

Through this project, I gained substantial experience in machine learning workflows (data cleaning, feature engineering, model tuning), best practices in DMPR data privacy considerations, and the importance of balancing business needs with predictive accuracy.

Cloud Project

Using AWS services—VPC, Security Groups, EC2, NAT Gateways, RDS, Application Load Balancer, Autoscaling Group, Certificate Manager, EFS, and Route 53—to host a WordPress Website. This project emphasizes cloud infrastructure design and DevOps practices, achieving scalability, security, and high availability for a production-ready environment.

API Project

This project automates the process of extracting schedule information from Excel files and converting it into events on Google Calendar using spaCy NLP models. Key steps include:

• Data Extraction: Reads Excel files, uses spaCy for entity recognition to identify event names, times, days, and dates in unstandardized formats.
• Event Grouping & Categorization: Groups recurring patterns (classes, meetings) through AI techniques, despite diverse formatting.
• Google Calendar Integration: Automatically creates and schedules events via the Google Calendar API, reducing manual data entry.
• Efficiency & Flexibility: Handles inconsistent input while syncing schedules quickly into Google Calendar.

Hackathon 24-hour Challenge

Collaborated on a hackathon project utilizing Pegasus, a text summarization model, to generate concise summaries from large datasets. Integrated AI-driven NLP techniques to automate document analysis, improving the speed and accuracy of content reviews. Delivered real-time data insights to end users via a custom AI pipeline.

Webpage For Class

Created a fully functional website using HTML, CSS, and JavaScript for class. Implemented responsive design principles and interactive elements to enhance user experience.