University Scholars Program, Sponsored presidential palace, This award recognizes faculty members who have conducted outstanding research and education, and their future achievements are highly anticipated. This award provides $15,000 per year for three years to enhance academic activities.
Subramanian Sankaranarayanan
Professor, Department of Mechanical and Industrial Engineering
UIC Institute of Technology
Years at UIC: 6
What themes or questions drive your research?
I develop physics-based artificial intelligence/machine learning approaches to accelerate materials discovery, design, and reliability. My group bridges the gap between electronic, atomic, and mesoscale models with multimodal experiments to understand how defects, interfaces, and metastable phases govern function, particularly in microelectronics, energy storage, thermal management, and catalysis.
How did you become interested in these fields of research?
That’s two things. Part of the appeal of understanding how simple atomic interactions determine complex behavior at the macroscale is the opportunity to combine high-performance computing and new AI techniques to turn rich experimental and synthetic materials datasets into predictive, designable models.
What kind of courses do you teach, and are there any topics that you particularly enjoy teaching?
I usually teach introductory thermodynamics. and Machine learning and data science for mechanical engineers. For the former, I will focus on the basics, as this subject cuts across nearly every engineering department (from engines and HVAC to batteries and manufacturing). The Machine Learning/Data course designs projects based on contemporary mechanical engineering and materials problems, allowing students to understand various aspects of an end-to-end machine learning pipeline, from data engineering and model selection to uncertainty and deployment, closely aligned with current industry demands.
What strategies help you balance teaching and research?
We aim to integrate them tightly. Class projects will use open datasets and problems from our laboratory and current literature. Students will learn how to contribute publishable results and develop design workflows using artificial intelligence and machine learning. In the classroom, the emphasis is first on problem formulation, that is, clarifying objectives, constraints, and assumptions, then selecting the appropriate tools (analytical models, simulations, experiments, or machine learning), and iterating on the data to arrive at a physically feasible solution. Milestone-driven mentoring, code and data sharing practices, and concise feedback loops keep both teaching and research moving in sync.
What advice would you give to students interested in a research career?
Build a strong foundation and communicate clearly both orally and in writing. Collaborate across disciplines and institutions. Be agile and learn and adapt. The most important thing is to be patient and persistent. Significant results often come after many careful iterations and checks.
