Satender Gunwal
PhD Student · Information and Decision Sciences · University of Illinois Chicago
I am a PhD student in the Information and Decision Sciences department at the College of Business Administration, University of Illinois Chicago, advised by Professor Selvaprabu Nadarajah. My research focuses on decision-centric AI and optimization for modernizing the electric grid and other high-stakes systems, where uncertainty, long planning horizons, and institutional constraints limit the effectiveness of traditional approaches. My current work emphasizes spatial reliability aware power system planning, the grid impacts of emerging demand such as large-scale data centers, and learning-augmented methods that translate complex data and simulations into actionable planning insights.
Prior to joining UIC, I worked as a pre-doctoral research associate at the Indian Institute of Management Ahmedabad, India, advised by Professor Ankur Sinha. There, my research focused on bilevel optimization formulations for hyperparameter optimization in machine learning.
Research Interests
Keywords: Decision-centric AI, Reliable Power Systems Planning, Planning under Uncertainty, Sustainability Analytics
Learning-augmented optimization for infrastructure and energy systems. I develop optimization frameworks for long-term generation and transmission planning that incorporate data-driven reliability signals while remaining computationally tractable. My work focuses on spatial reliability-aware power systems planning, the integration of machine learning surrogates into planning models, and evaluating the grid impacts of emerging large-scale demand such as data centers.
Decision-centric AI for sustainability and policy planning. I study how machine learning and large language models can support sustainability decision-making in public-sector and resource-constrained settings. This includes developing frameworks to structure, benchmark, and compare municipal sustainability plans, enabling peer learning, financing insights, and scalable pathways from local action to global sustainability goals.
Working Papers
We study long-term generation and transmission expansion planning where deterministic investment models are evaluated ex post for probabilistic resource adequacy, a practice that can overlook spatial reliability. We show that a machine learning–based reliability feedback can uncover expansion plans that dramatically reduce outages and rebalance reliability across regions at minimal additional cost, without introducing stochastic planning complexity.
Published Papers
We propose a linear programming–based “hyper local search” method for hyperparameter tuning. By formulating hyperparameter optimization as a bilevel program, the approach identifies a local descent direction to improve validation loss and simultaneously fine-tunes continuous hyperparameters and model parameters near an initial trained model. We demonstrate the approach across regression, machine learning, and deep learning tasks.
Work In-Progress
This project develops a value-based framework for evaluating water use at hyperscale data centers by linking regional hydrology, infrastructure constraints, and operational decisions to the opportunity cost of water use. We introduce a value-adjusted water footprint metric and a decision-factor mapping to identify best-in-class water management practices and benchmark current industry performance across Midwestern regions.
This project studies how AI-enabled analytics and large language models can be used to structure, benchmark, and compare municipal sustainability plans across regions. The objective is to support data-driven peer learning, identify financing and policy gaps, and enable scalable pathways from local environmental action to global sustainability goals.
This project uses large language models to analyze corporate nature- and biodiversity-related disclosures in the CDP questionnaire. The objective is to systematically classify goals, commitments, and targets, assess alignment with material impacts and dependencies using ENCORE and Global Biodiversity Framework (GBF) categories, and identify sector-level reporting patterns and gaps. The analysis aims to benchmark common practices, highlight opportunities to improve measurability and time-bound target setting, and support data-driven guidance for companies advancing nature-related disclosure and target-setting practices.
Teaching
- Course Development Assistant (with Professor Selvaprabu Nadarajah): Co-developed two courses at the University of Illinois Chicago — IDS 509: Foundations of Analytics and AI for Supply Chain and Operations Management and IDS 494: Data-Driven Decisions for Sustainable Business.
| Role | Course | Institution | Term(s) |
|---|---|---|---|
| Teaching Assistant | IDS 509: Foundations of Analytics and AI for Supply Chain and Operations Management | University of Illinois Chicago | Fall 2024, Spring 2025 |
| Teaching Assistant | IDS 494: Data-Driven Decisions for Sustainable Business | University of Illinois Chicago | Fall 2024, Spring 2025 |
Talks & Awards
Invited Talks
- Commodity and Energy Markets Association (CEMA) Annual Meeting, Houston, TX · 2025
- INFORMS Annual Meeting, ENRE Electricity, Atlanta, GA · 2025
- Annual POMS Conference, Industry Studies and Public Policy, Reno, NV · 2026
- Sustainability Research and Innovation (SRI) Congress (Workshop Session), Chicago, IL · 2025
- Biodiversity + Nature in Energy and Transportation (BNEAT) Meeting, Chesterton, IN · 2025
Awards
- Graduate Student Council, University of Illinois Chicago · 2025
- POMS College of Sustainable Operations (CSO), for the POMS conference · 2026
Curriculum Vitae
If the embed doesn’t load in your browser, use the download link.