✦ |
Data Scientist pursuing a Master's at the University of Maryland. I build intelligent systems — from computer vision pipelines on AWS to LLM-powered research tools — that turn messy real-world data into decisions that matter.
Built RockyBot — a LangChain-powered retrieval pipeline analyzing insights from 5,000+ financial articles using OpenAI embeddings and FAISS for sub-second semantic search. Achieved 95% retrieval accuracy.
Built an ETL pipeline in SQL Server processing 1M+ telecom records, improving query performance by 45%. Integrated a 92% accurate Random Forest model into an interactive Power BI dashboard.
Created a Power BI dashboard analyzing $30M in total sales. Tracked YTD, PYTD, GP%, and Sales Quantity across countries and product types, improving report speed by 25%.
End-to-end ML pipeline achieving 92% accuracy. Engineered custom logging, exception handling, and containerized with Docker for deployment across 10+ machines.
AI-driven YOLO model detecting animals in real-time with 85% accuracy. Reduced false positives by 30% and integrated with alarms and ultrasonic systems, cutting crop damage by 40%.
Anomaly detection system using PySpark, SQL, and AWS achieving 98% accuracy. Processed large-scale financial transaction data to flag suspicious patterns in real time.