My research interest lies in applying Data Science and Machine Learning techniques to the field of Ab-Initio Computational Materials Science. Currently in my 4th year of PhD program also pursuing a Doctoral Certification in Data Science.
Tools: TensorFlow, PyTorch, VASP, ATAT, Python, C++
Working on computer calculation of phase diagrams and generation of atomic potentials under Dr. Axel Van De Walle.
Implemented a DenseNet architecture and trained from scratch with data augmentation and hyperparameter tuning to recognize handwritten bengali characters. Secured an accuracy of 92% with 1 GPU and 100 epochs of training.
Fine-tuned a GPT-2 language model to generate Amazon reviews like text conditioned on keyword and rating. Also fine-tuned a PPLM model for better sentiment control.
Calculated from 1st principles, the phase diagram for a Iridium-Ruthenium binary alloy using sqs2tdb
Mesoporous lipid-silica nanohybrids for folate-targeted drug-resistant ovarian cancer
Sayan Samanta, Lina Pradhan, D. Bahadur, New Journal of Chemistry, 42 (4), 2804-2814; 2018
Interfacing ab initio calculations, Calphad models, thermodynamic databases, web interfaces and visualization tools
The summer school discussed both in theory and hands-on computational techniques in materials science across different length-scales. Topics covered included: Density Functional Theory, Cluster Expansion, Molecular Dynamics, Dislocation Dynamics, Micromechanics and Machine Learning
Structural characterization of mesoporous copper oxide as prepared by high energy ball milling from low angle x‐ray scattering experiments
Thermodynamic modeling of solid solubility change during mechanical alloying of Cu-Co system
Studied the principles of ergonomic and affordable design within the context of rural India. Participated in a bunch of maker‐space exercises such as :
Synthesis & Structural Characterisation of Mesoporous Copper Oxide as prepared by High Energy Ball Milling from Low Angle X-Ray Scattering.