-Extending “Cut and Columnize” Mixed Integer Programming strategy beyond Multicommodity Flow Problems with Service Requirements to Constrained Vehicle Routing and Scheduling Problems

CS395T Introduction to Ethical AI

-Conducted research to develop motion-planning algorithms, developed stochastic programming formulation to implement probabilistic roadmap planner in robot configuration space for collision-checking under uncertainty

-Implemented an economic dispatch and unit commitment Mixed Integer Programming model to optimize EV charging schedule to maximize margin on TOU rates-Conducted scenario analyses of Vehicle-to-Grid Integration in an Electric Vehicle-based Virtual Power Plant to determine locational feasibility of power plant under various utilities’ electricity tariff structures

CE311S Statistics for Civil Engineers

-Led the development of an integrated process flow model to analyze the viability of coal plants for Carbon Capture and Storage (CCS) integration; created a carbon capture power system module in Python to study optimal economic interactions between the base and auxiliary plant in a power system-Managed and integrated a large MySQL database of United States power plants across multiple US agencies and organizations-Collaborated with multiple national labs and consultancies to work on providing rapid predictive analytics for storage reservoir response to real-time change in subsurface operations using physics-based ML models-Developed an economic dispatch model with water consumption analytics using Pyomo to study power system water shortages in Texas; presented at 2019 ASME Power Conference and published at 2019 International Water Conference-Created an internal dashboard in Tableau to identify coal plants for carbon capture based on technical and geospatial constraints for the Department of Energy (DoE)-Designed a data science pipeline to analyze water pollution from Northern Appalachian power plants using pollutant discharge permit data-Co-authored DoE reports analyzing reliability and economic factors of anthropogenic CO₂ sources for enhanced oil recovery-Worked across multiple teams to conduct data analysis, visualizations and create Python-based applications

-Developed an open-source and interpretable spatio-temporal energy demand allocation model for mainland US-Wrangled socio-economic indicators from multiple energy and census bureau datasets for model parameter estimation-Conducted git integration of database with company’s in-house PUDL (Public Utility Data Liberation) project

-Developed day-ahead load (mean average percentage error <15%) and solar generation forecasting model (mean absolute scaled error <0.1) using past load, weather and time features-Developed model utilizing deep network with convolutional and recurrent LSTM neural network layers.

-Automated CMU’s Power Sector Carbon Index website to categorize and visualize quarterly US emissions and power generation by fuel type in Python’s Jupyter environment-Creating map and time series visualizations for emissions and power generation, categorized by NERC regions

- Emerging Energy Policies 19-621 (Fall '17)- Energy Innovation and Entrepreneurship 19-627 (Fall '17)- Energy Demand and Utilization 39-611 (Spring '18)- Energy Economics and Policy 39-612 (Spring '18)- Machine Learning 10-601 (Peer Tutor) (Spring '18)

-Assisted in evaluating Expression of Interest of companies for EESL’s business opportunities outside India-Analyzed financial performance of EESL and compared with other Energy Service Companies

-Conducted A/B testing, text and image features analysis; content creation for the International section of news to maximize time spent on the news application

-Implemented MATLAB validation for IGBT switching characteristics, devised a scheme using DPlot to systematically categorize and group the raw test results-Supervised testing of industrial IGBT double pulse test and traction motor validation