UMass Amherst – NSF PRISM Research (2025–2026): Built and evaluated ML pipelines for a federally funded research project, designing evaluation frameworks to benchmark model performance and compare retrieval strategies across large document corpora. UMass Amherst – Summer REU (2025): Analyzed 22 years of U.S. tariff and trade flow data using statistical methods including Pearson/Spearman correlation, Granger causality, Mutual Information, and Random Forests. Engineered a hybrid VMD-LSTM forecasting model that outperformed baseline models by 27%. Co-authored a research paper on the findings. Sky Governance (2024): Built data pipelines using Python, Pandas, and SQL across AWS, Azure, and GCP for real-time anomaly detection and compliance monitoring on enterprise governance data. Accurate Industrial Controls (2023): Trained and validated ML models (CNN, Decision Tree) on real-time industrial sensor data using TensorFlow, Scikit-learn, and SciPy, improving robotic arm efficiency by 20%. Clause – HackUMass 2025 (Won Best Use of AI): Built an AI-powered legal document analysis platform using RAG pipelines, PII redaction, and LLM-grounded analysis to extract insights from leases and medical bills at scale.…UMass Amherst – NSF PRISM Research (2025–2026): Built and evaluated ML pipelines for a federally funded research project, designing evaluation frameworks to benchmark model performance and compare retrieval strategies across large document corpora. UMass Amherst – Summer REU (2025): Analyzed 22 years of U.S. tariff and trade flow data using statistical methods including Pearson/Spearman correlation, Granger causality, Mutual Information, and Random Forests. Engineered a hybrid VMD-LSTM forecasting model that outperformed baseline models by 27%. Co-authored a research paper on the findings. Sky Governance (2024): Built data pipelines using Python, Pandas, and SQL across AWS, Azure, and GCP for real-time anomaly detection and compliance monitoring on enterprise governance data. Accurate Industrial Controls (2023): Trained and validated ML models (CNN, Decision Tree) on real-time industrial sensor data using TensorFlow, Scikit-learn, and SciPy, improving robotic arm efficiency by 20%. Clause – HackUMass 2025 (Won Best Use of AI): Built an AI-powered legal document analysis platform using RAG pipelines, PII redaction, and LLM-grounded analysis to extract insights from leases and medical bills at scale.WWWW…