Hi, I'm

Prosun Datta

PIML Researcher | Founder & Executive Director, Dynexly
Youth-led research collective. Physics meets machine learning.
Currently →
Physics-Informed ML Deep Learning Computational Physics Open Source Optics & Microgrids
00. About

Who I Am

My name is Prosun Datta, and I am a researcher operating at the intersection of Computational Physics & Machine Learning. Based in Sylhet, Bangladesh, my work focuses on the structural resilience of critical infrastructure, specifically developing forecasting models for energy systems in volatile environments.

In March 2026, I founded Dynexly (Dynamic Neural Exploration Lab), a youth-led research collective for individuals aged 14–22. We operate as a high-rigor, open-source initiative where every project targets peer-reviewed standards, bridging the gap between advanced ML research and youth access in South Asia.

My current research specialization involves Physics-Informed Machine Learning (PIML). By embedding physical laws directly into neural networks, I am building models that provide superior reliability for renewable energy forecasting in hospital microgrids, particularly within the flood-prone monsoon corridors of Bangladesh.

Looking ahead, I am navigating my HSC 2027 candidacy as a foundational period for advanced academic rigor. I am actively seeking research collaborations and faculty mentorship in applied mathematics and PIML, aiming to leverage my upcoming undergraduate transition to scale mathematical frameworks that ensure technological stability for communities in the Global South.

Discover the human side Beyond the Lab
01. Research

Flagship Research

Physics-Informed Machine Learning applied to real infrastructure problems.

02. Projects

Technical Work

Computational projects in ML and physics simulation.

03a. Skills

Expertise

03b. Education

Academic

04. Writing

Articles

Notes on machine learning, physics, and scientific thinking.

05. Contact

Get In Touch

Open to research collaborations, mentorship, and joint work in PIML and computational physics.