Sangram Keshari Patro

Sangram Keshari Patro Linkedin || Github || Portfolio ||-|| - SKILLS SUMMARY Programming Languages & Tools Aspiring data scientist with a background in computational physics and practical experience in machine learning, predictive modeling, and time series forecasting. • Python • C++ • Matlab • Latex • SQL • Fortran • Microsoft Office • Shellscripting PROJECTS • Radware • Arduino • Origin • Markdown Data Science tools: Forecasting Wine Sales - ABC Estate Wines Mar. 2025 • Scikit-learn • Statsmodels • Matplotlib • Forecasted 12-month wine sales across categories using Exponential Smoothing • Pandas • Numpy • Tableau • Seaborn and ARIMA on 30+ years of data(RMSE: $9,334), uncovering Q4 seasonality • Plotly • Shap • KNIME and long-term sales decline to inform rebranding and premium festive marketing strategies. EDUCATION THE UNIVERSITY OF TEXAS AT AUSTIN (Online) PGP in Data Science and Business Analytics Expected: Aug 2025 Current GPA: 4.1 / 4.33 UM-DAE Centre for Excellence in Basic Sciences, Mumbai Bsc. and Msc. in Physics Apr 2024 | Mumbai, India Cum. GPA: 3.2 / 4 COURSEWORK UNDERGRADUATE Analyzing Sales Trends and Customer Feedback - New Wheels Feb. 2025 • Analyzed unstructured sales and customer data using advanced SQL, identifying a 35.8% order drop, 47% rise in negative ratings, and 1.16 point satisfaction decline to inform customer retention, pricing, and logistics strategies. Customer Segmentation Using Unsupervised Learning - AllLife Bank Jan. 2025 • Segmented 600+ credit card users via K-Means, Hierarchical Clustering, and PCA into 3 behavioral groups, revealing a 2.5x higher credit limit and 1.8x higher avg. card counts in digital users, enabling targeted retention & upsell strategies. Optimization of the visa approval process - EasyVisa Dec. 2024 • Built and tuned an XGBoost classifier(F1: 0.825) to optimize visa certification predictions across 25K+ cases; Used SHAP analysis to quantify feature impact, revealing that every $10K wage increase raised visa approval odds by ~1.3%. Booking Cancellation Prediction - INN Hotels Nov. 2024 • Developed Logistic Regression & pruned Decision Tree models(F1: 0.80) to Computational Physics predict hotel booking cancellations on 36K+ records & identified actionable drivers Numerical Methods such as lead time, special requests; projected to reduce cancellations by 22%. Linear Algebra and Differential Equations Mathematical Physics OTT Data Analysis - ShowTime Statistical and Computational Techniques Oct. 2024 Statistical Mechanics • Built a linear regression model (Adj. R²: 77.4%) to predict OTT first-day views Biostatistics from various factors, revealing 6.5M more visitors raise views by 1M, while sameday sports events cut views by 63K. PGP Predictive Modeling EXPERIENCES Machine Learning Basics Tata Institute of Fundamental Research | Graduate Research Intern Machine Learning Advanced Jun 2023 - Dec 2023 | Mumbai, IN Inferential Statistics • Developed a Python program using chi-square minimization to fit time spectra Unsupervised & Supervised Learning with a 5-parameter Gaussian+exponential model, improving lifetime extraction of Time Series Forecasting ¹³³Cs; observed better fit and parameter stability beyond two half-lives. ACHIEVEMENTS Tata Institute of Fundamental Research | Summer Research Intern | Mumbai, IN Compared HPGe clover detector efficiencies with/without CsI(Tl) detectors; visualized results using Python, Excel, Matplotlib, and Radware. Modeled cranked shell-model in C++ with higher-order deformations; applied curve fitting and statistical analysis to match experimental data and predict unexplored regions. May 2022 - Jul 2022 DATA SCIENCE COMPETITIONS • • Top 4/1K Analytics Vidhya Global AI Battlefield • Top 3 in Great Learning Hackathon DISHA scholar(Aug 2019 - Apr 2024) by Department of Atomic Energy (Awarded to Top 50/40k applicants) NSTSE – 2019 AIR -104 conducted by Unified Council. NEST - 2019 AIR -70 PUBLICATIONS A. Research Contributor, A Book on Physics and the Cities of India (Ed. Himanshu K. Pandey, IIP, 2021) • Researched Tensegrity structures, Lingaraj Temple physics, and Samanta Chandra Sekhar’s astronomical instruments under Prof. Vijay A. Singh; contributed a section to a published book (ISBN:-).