Patrick Moon

Built and maintained Databricks-based ETL and analytics pipelines using PySpark and SQL, embedding data quality checks to improve completeness, validity, and consistency metrics to 90%+ across enterprise datasets. Managed an enterprise commercial decision model used by bp SVP leadership and partner VPs, enabling multi-scenario evaluation of offshore drilling and completion programs valued at ~$300

Built and deployed an AI agent for natural language querying using LangChain and Azure OpenAI, enabling non-technical users to access enterprise datasets and reducing reliance on ad hoc SQL queries by ~40%. Applied supervised classification models (XGBoost, Random Forest) with rule-based validation to identify missing metadata in Azure DevOps, enabling automated remediation for ~50% of incomplete

Developed Convolutional neural network-based computer vision classification models to identify bee species from wing images, achieving ~95% test accuracy across 50+ visually similar species. Built image preprocessing pipelines including perspective correction, cropping, and normalization to standardize lab and field images, improving model robustness across heterogeneous data sources.

Analyzed large-scale phonetic transcription datasets from linguistic experiments, applying NLP preprocessing and feature extraction in R to transform unstructured text into structured inputs for statistical modeling. Conducted exploratory and inferential statistical analyses to identify phonetic and temporal patterns across speaker groups, supporting hypothesis testing and quantitative validation

Grade: 3.75

Graduated with a Bachelor of Science in Statistics and Data Science, achieving a cumulative GPA of 3.75.