● Designed and built an AI-driven document tagging system to classify documents by category and metadata.

● Implemented and compared multiple NLP and ML approaches, including DistilBERT, ModernBERT, Bag-of-Words,

and SVM, to optimize classification accuracy.

● Accelerated experimentation and training workflows using Databricks, Weights & Biases, and Optuna for

hyperparameter optimization and experiment tracking.

● Evaluated model performance using precision, recall, and F1-score metrics, achieving strong and reliable

classification results.

● Designed and implemented a regex-driven tagging engine to compare deterministic lexical results against AI-based document classification models