Research
Since my third year of undergraduate studies, I have been actively engaged in research across several domains, including molecular dynamics, machine learning, control systems, and mechatronics. A defining aspect of my work has been the integration of deep learning techniques into mechanical engineering applications, an area with significant untapped potential. My research is focused on pioneering the use of these advanced methodologies in fields where they have yet to be fully explored.
material behavior
Metal solidification behavior, crystalline structure growth, crack propagation in 2D materials, atomistic scale properties of alloys.
machine learning
Convolutional Neural Networks, Reinforcement Learning, Time Series Analysis, Hybrid Deep Learning Model, Machine Vision.
additive manufacturing
Laser Powder Bed Fusion process, Effect of process parameters on the mechanical properties of the manufactured parts.
ROBOTICS AND CONTROL SYSTEms
Machine Vision, Robotic Arm manipulation, PID Control, Reinforcement Learning based control system.
Publications and Conference Procedings
Accepted
Automated Waste Sorting using Deep Learning and Robotic Manipulation: A Comprehensive Approach
Taaha Md Tanvir Hossain, Dastagir Rafi B, Haque Md Amin, Muhit M Abrar
International Conference on Mechanical, Industrial and Materials Engineering (ICMIME), 2024
Under Review
Prediction of Crystalline Structure Evolution During Solidification of Aluminum at Varying Cooling Rates Using a Hybrid Neural Network Model
Dastagir Rafi B, Chanda Shorup, Chowdhury Farsia K, Chowdhury Shahereen, Rahman K Arafat
Journal Publication
Manuscript in Progress
An AI-Driven Smartphone Solution Enhancing Accessibility for Visually Impaired Users
Dastagir Rafi B, J. T. Jami, Chanda Shorup, M. Rahman, K. Dey, F. Hafiz, M. M. Rahman, M. Qureshi and M. M. Chowdhury,
Manuscript to be submitted.
Research Experience
Research Assistantship
Department of Mechanical Engineering, BUET
Under the Guidance of Dr. Shahereen Chowdhury, Assistant Professor, Dept. of Mechanical Engineering, BUET
My current research involves the application of 3D convolutional neural networks in predicting material properties at atomistic scale.
SUPREME 2024 - Summer Program
CIMPI LAB, Washington State University
Under the Guidance of Dr. Satyajit Mojumder, Assistant Professor, Washington State University
Participated in a summer research mentorship program under the supervision of Dr. Satyajit Mojumder, Assistant Professor at Washington State University, where my task was developing a reinforcement learning-based control system for the laser powder bed fusion process while also writing an analytical solver in Python.
Undergraduate Thesis
Implementation of Hybrid Conv1D-LSTM model in predicting the crystal structure evolution of Aluminum at different cooling rates.
Supervisor: Dr. Shahereen Chowdhury, Assistant Professor, Dept. of Mechanical Engineering, BUET
We have been working on training and tuning multiple deep learning models to predict the non-linear behavior in evolution of crystal structures in Aluminum at different cooling rates. Additionally, I have devised and implemented an automated procedure for conducting a large number of molecular dynamics simulations, extracting analysis data from OVITO, and assembling a comprehensive dataset.
Laser Powder Bed Fusion
Currently working on an atomistic simulation of Laser Powder Bed Fusion of Invar 36 to examine its mechanical properties at different laser parameters i.e. laser power and speed.
Research in the field of AI
Trained and optimized an image classifier applying explainable AI to assist vision-impaired users in identifying rapid test kit results via mobile application.