Soumya Diddi

I’m an ECE (Honors) graduate specialized in Data Science & Big Data Analytics, with a strong quantitative and computer-science foundation. I apply machine learning and statistical modeling—especially XGBoost with SHAP explainability and feature engineering—to turn large datasets (100K–500K+ records) into clear, actionable insights for both technical and business stakeholders.

In my Summer Research Intern role at IIT Tirupati, I built Python data pipelines to extract, profile, and clean 100K+ records, adding data-quality and validation checks to optimize performance. My pipeline trade-off improvements reduced per-frame processing time by ~18% with minimal accuracy loss, and my technical report was adopted by the lab for follow-on research. I’ve also delivered analytics products like a YouTube Trending Content Intelligence platform (200K+ videos, 82%+ viral-probability classification, and SHAP-driven driver explanations) and a Tech Product Reviews dashboard (ETL for 500K+ reviews with <1% data loss, advanced SQL, and Tableau/refresh automation).