Aditya Sai Pranith Ayapilla

Brief intro... I’m Pranith, a postdoctoral researcher at the Institute of Science Tokyo (Onishi Laboratory). In plain language, I use computers and math to try to understand the physics of how fluids flow. But more formally, my work sits in the intersection of fluid dynamics and scientific machine learning, with current interests in ML-based numerical weather prediction, data assimilation and urban micrometeorology.

Quick background... I was born in Visakhapatnam (Vizag), a coastal city on the east coast of India. I completed my undergraduate studies in Mechanical and Aerospace Engineering at IIT Hyderabad (2020), and then moved to Sendai, Japan, for graduate studies at Tohoku University under Prof. Yuji Hattori, supported by a fully funded JICA scholarship. I received my PhD from Tohoku University in 2025.

Current research motivation... Over the past few centuries, scientists have developed a deep understanding of the physical world using three main approaches: experiments, theoretical analysis, and computer simulations. In recent years, advances in artificial intelligence (AI) have begun to influence many scientific fields, including fluid mechanics, which is central to problems such as weather prediction, aerospace engineering, and environmental sustainability. Many fluid dynamical systems, especially atmospheric flows, are highly complex, involving interactions across multiple scales and often exhibiting chaotic behavior. Because of this complexity, theoretical analysis is often difficult and high-fidelity numerical simulations can be extremely computationally expensive, making real-time prediction challenging. This is precisely where AI, in the form of scientific machine learning (SciML), has been time and again shown to be very useful. SciML can learn from expensive high-resolution simulations and later reconstruct high-resolution fields much faster from cheaper low-fidelity information. 

My current line of research focuses on urban micrometeorology, which refers to weather behavior at the scale of cities and neighborhoods. I 'm working to combine methods of deep learning, generative AI, etc. with my background in fluid physics to develop computationally efficient, data-driven models that enable fast and accurate urban micrometeorological modeling and prediction. This is becoming increasingly important as rapid urbanization and climate change lead to weather patterns in cities that differ significantly from those in surrounding regions.

Previous work... Before my current work, in my PhD research, under the supervision of Hattori-sensei, I explored a more classical fluid-dynamics problem with more of a first-principled approach: a specific type of instability called the triangular instability, which can cause strained vortices to become unstable. Using theoretical analysis and numerical simulations, I showed that this instability can arise in Batchelor vortices (a widely used model for vortices in wake flows), and elucidated its characteristics, with potential relevance to the modeling and control of wind-turbine wakes.

• Click here for a more detailed overview of my research experience and interests.

• Click here for my curriculum vitae.

• Profile and social links: Google Scholar, ResearchGate, LinkedIn.

• The codes I developed/contributed to can be found on my Github.