Stanton Hamilton

- Aided the training of large language models (LLMs) through annotation of model responses, prompt development, and coding corrections.

- Used Python to help train models to write quality code.

- Wrote human-readable summaries of problems.

- Solved complex data problems, like data generation, data cleaning, data visualization, and machine learning modeling.

- Developed models to support travel nurse staff hires for national hospital systems. These machine learning models included segmentation, time series prediction, and binary classification, to predict nursing candidate success and identify seasonal staffing trends. Used methods such as random forests, XGBoost, Regression, DBSCAN, K-Means, NLP, and SARIMAX.

- Utilized NLP processes to identify similar reports to reduce labor requirements for a company-wide data and report migration.

- Created functional dashboards in Domo for stakeholders to monitor metrics and identify trends. I was then integral in the migration of company-wide reports from Domo to Power-BI.

- Worked to ensure data quality in processing and reporting. I did this by running quality checks in Open Meta Data, and working to address any issues that arose.

- Presented analysis results to stakeholders including executives.

- Performed inventory optimization analyses for both website and business to business sales between warehouses, based on sales trends and inventory needs.

- Created a linear regression model to predict trailer unload labor times of new product, with the goal of labor time optimization.

- Performed a financial analysis to support a decision to increase warehouse capacity by 50%.

- Created tools in Excel and Python, including regression models, to estimate time and labor costs for various warehouse processes.

- Developed advanced SQL Queries to return data regarding past sales and company trends.

- Used Domo to create data visualizations and dashboards to aid visibility of company data.

- Identified duplicate purchases of digital publications from multiple journals to reduce future costs for the library in maintaining its records.

• Scraped publication metadata from multiple journal’s APIs.
• Created a SQL database to store this information for later analysis.
• Developed a binary classification model using XGBoost to identify duplicate records due to dissimilar data. I achieved an 89% accuracy rate.

- Developed data visualizations, reports and presentations for university administrators.