Machine Learning & AI

A Python-based AI chatbot designed to improve truck dispatching efficiency and compliance.

• Improved compliance tracking through intelligent automation
• Connected with Google Sheets using Python, Pandas, and Seaborn for analysis and visualization
• Structured JSON and optimized query logic to enhance AI response accuracy and performance

A proof-of-concept machine learning model built to increase efficiency for 100+ phone dispatchers.

• Python-based ML model
• Designed to optimize dispatcher workflows
• Part of my role at BYU-I Support Center
• Integrated into existing support center infrastructure

An exploration of creativity and coherence in AI-generated writing using vastly different literary sources.

• Built to explore creativity and coherence in AI-generated writing using vastly different literary sources
• Trained an RNN to generate coherent text based on a blend of Shrek scripts and Dr. Seuss stories
• Focused on NLP model fine-tuning for narrative style and syntactic coherence

A machine learning project focused on improving real estate price estimation.

• Tackled the problem of price estimation inconsistencies in regional real estate data
• Built an XGBoost model to predict home prices; reduced RMSE from 200K to 80K
• Discovered bedroom count as a key predictor through feature analysis

A neural network approach to used car marketplace valuation.

• Created to simulate used car marketplace valuation using historical feature-rich datasets
• Developed a neural network model with log-transformed targets to improve accuracy
• Validated model improvements with lower prediction variance

A business-focused forecasting model for bike rental inventory planning.

• Targeted business needs to better forecast rental spikes for inventory and maintenance planning
• Preprocessed and optimized data to train a high-accuracy neural network
• Tuned hyperparameters and model architecture to improve demand forecasts

A customer churn prediction model to help banks retain customers.

• Designed to help banks proactively identify customers at risk of leaving based on behavioral patterns
• Created a random forest model to predict customer churn, boosting recall from 25% to 69%
• Focused on customer segmentation and actionable ML-driven targeting