M Arshad Zahangir Chowdhury

My work builds on classical mechanical engineering principles, integrating interpretable AI techniques to support advanced analysis, monitoring, and system design.

Rensselaer Polytechnic Institute, Troy, NY 12180.

chowdm6ATrpi.edu.


2023 -
I started my academic career at RPI. I teach primarily Elements of Mechanical Design. I also teach Machine Learning for Engineering, Fluid Dynamics Laboratory, Thermal and Fluids Laboratory, and Engineering Economics.
2018 - 2023
I received my PhD from Rensselaer Polytechnic Institute in Mechanical Engineering advised by Prof. Matthew A. Oehlschlaeger.
2021 - 2022
I was a research aide at the Data Science and Learning Division at Argonne National Laboratory.
2016 - 2018
I obtained my Master's degree from then Indiana University Purdue University Indianapolis (present day Purdue University in Indianapolis), advised by Prof. M. Razi Nalim.
2014 - 2016
I served as a lecturer of mechanical engineering at Ahsanullah University of Science and Technology, Dhaka, Bangladesh.
2009 - 2014
I majored in Mechanical Engineering with a Bachelor of Science in Mechanical Engineering at Bangladesh University of Engineering and Technology.
Research Vision

My research is rooted in the foundational disciplines of mechanical engineering, solid mechanics, fluid dynamics, heat transfer, and system dynamics, with a focus on solving problems that demand both physical insight and computational innovation. I view artificial intelligence not as a replacement for traditional engineering analysis, but as a powerful set of tools that when used judiciously can supplement and extend classical methods. Many real-world engineering systems; such as those involved in energy infrastructure, robotics, manufacturing, and environmental sensing operate in conditions that are too complex, uncertain, or data-rich to be addressed by purely analytical or first-principles models alone. In this context, my vision is to develop hybrid approaches that combine mechanistic models with data-driven techniques, enabling engineers to make better decisions in design, diagnosis, and control. I am particularly interested in using interpretable machine learning to identify patterns in data, enhance reduced-order models, and support real-time monitoring of mechanical systems. My long-term goal is to build tools that are computationally efficient, physically consistent, and transparent; tools that preserve the trust and reliability of classical engineering approaches while benefiting from the adaptability and scalability of AI. Through interdisciplinary collaboration and a strong foundation in engineering fundamentals, I aim to advance methods that empower mechanical engineers to address increasingly complex challenges without losing sight of the principles that define the discipline.

Prior Research Work

See below for some of my detail work during and post-PhD and one of the review I wrote for further details.

Deep Learning for Gas Sensing via Infrared Spectroscopy

In this project, A deep learning model is trained on simulated absorbance spectra to recognize infrared spectra of up to three component mixture in a broad 400-4000 1/cm wavenumber range. The proposed model is capable of highlighting important frequencies for multi-gas sensing and probable cause for misclassification within a user defined frequency range.

TSMC-Net : Deep Learning Mixture Component Identification

In this project, A deep learning model is trained on simulated absorbance spectra to recognize THz spectra of up to eight component mixture in 220-330 GHz frequency range. The workflow proposed is capable of highlighting important frequencies for multi-gas sensing and probable cause for misclassification by the model.

VOC-Net: Deep Learning Volatile Organic Compounds Classifier

In this project, A deep learning 1D-CNN based classifier is trained on simulated absorbance spectra to recognize VOC THz spectra in 220-330 GHz frequency range.

IR Spectra Classification

In this project, SVM models are trained on simulated absorbance spectra to recognize IR spectra in a broad frequency range.

X-Ray Fluoroscence Imaging ROI-Finder

Code for intelligently guiding X-ray fluorescence scans to regions of interest using a suite of ML-based clustering algorithms.

Rotational Spectra Classification

This code trains on simulated absorbance spectra to recognize measured THz spectra in the 220-330 GHz frequency range using several machine learning methods.

Students

Supervised Undergraduate Research Projects (URPs)

Ian Oehlschlaeger, Machine learning for sports analytics.

Michael Ahn, Generative design of machine elements.

Runhan Gu, Classify and autoencode gas spectral data to make a gas sensing digital twin.

Teaching

At Rensselaer Polytechnic Institute, Troy, New York

MANE 4030: Elements of Mechanical Design [Sum'23, Fall'23, Spring'24, Sum'24, Fall'24, Spring'25, Sum'25]

MANE 4962: Machine Learning for Engineering [Spring'23,Spring'24, Spring'25]

MANE 4910: Fluid Dynamics Laboratory

MANE 4740: Thermal and Fluids Engineering Laboratory

ENGR 4760: Engineering Economics

At Ahsanullah University of Science and Technology, Dhaka, Bangladesh

ME 201: Elements of Theory of Machine and Machine Design

ME 2201: Mechanics of Materials

ME 2110/ME 102/ME114: Mechanical Engineering Drawing I

ME 3204: Control Engineering Sessional

ME 3206: Heat Transfer Sessional

ME 4102: Heat Engines Sessional

ME 302: Fundamentals of Mechanical Engineering Sessional