Debojyoti Das

1. Liang, R., Das, D., & Bakhtiiari, A. (2021). Protein confinement fine-tunes aggregation-induced emission in human serum albumin. Physical Chemistry Chemical Physics, 23(46), 26263-26272. Https://doi.org/10.1039/D1CP04577F

2. Sarka, J., Das, D., & Poirier, B. (2021). Calculation of rovibrational eigenstates of H3+ using ScalIT. AIP Advances, 11(4), 045033. Https://doi.org/10.1063/5.0047823

3. Project-1, 4, 5, publications under review in peer-reviewed journals

4. Presented talk on “Recent development in END simulation for ICT reaction” in ACS Fall 2023 meeting and, presented poster in TACC 2019.

• Machine Learning-Driven ADMET Prediction for Doxorubicin and its Derivatives: Developed machine learning models for ADMET prediction of Doxorubicin and its derivatives using a dataset of 2700 molecules, generated by using RDKit
• Employed Random Forest for a 92% accuracy, and enhanced drug research efficiency through molecular docking and dynamics simulations with GROMACS.
• EGFR Kinase Virtual Screening and Drug Discovery: Orchestrate an automated pipeline for virtual screening, enhancing EGFR kinase inhibitor selection
• Developed pharmacophore models through molecular docking, and refined predictive accuracy using ML techniques like VAEs, GANs, and ANNs, achieving a 96% success rate.
• Aggregation-Induced Emission Fine-Tuning through Protein Confinement: Pioneered ab initio simulations to decode AIE in biomolecules, developed a novel electronic method, and uncovered proteins' role in decay pathways using CHARMM in OpenMM, enhancing AIE applications in bioimaging and diagnostics.
• Ion-Cancer Therapy (ICT) Reactions Optimization Through Machine Learning: Implemented advanced ML algorithms, including RNN, CNN, and VAE within PyTorch and scipy frameworks, to refine ICT reaction predictions
• Achieved notable precision with a 0.67 error for rainbow angles and 0.56 for impact parameters using ANNs, demonstrating improved method efficacy and computational cost reduction.
• Enhancing SpaceX Launch Success Predictions through Machine Learning": Developed decision tree classifier, achieving predictive accuracy for first-stage landings and contributing to an increased launch success rate from 2013-2020.
• Predictive Modeling of Reaction-Types in Ion-Molecule Collisions Using Advanced RNNs: Pioneered the use of RNNs for classifying p + C2H4 collisions, overcoming quantum challenges and achieving a standout 98.23% accuracy
• This innovation bridged quantum chemistry and machine learning, establishing new standards for ICT reactions.
• Master’s Thesis: Ro-vibrational Spectroscopy of H3+
• Utilized ScalIT to calculate H3+'s energy levels, achieving convergence to 10-4 cm-1 for vibrational states up to J = 46
• Validated results with existing research, contributing to molecular astrophysics and theoretical chemistry knowledge.
• Developed comprehensive reports on research progress, prepared grant proposals, keeping PI informed and engaged.

• Collaborated with teachers for lesson preparation by preparing materials and setting up equipment.
• Reviewed lesson material with students individually or in small groups.
• Handled class records for attendance, assignment grades and course participation scores.

• Maintained database systems to track and analyze operational data.
• Developed organizational skills through managing multiple tasks simultaneously while adhering to strict deadlines.
• Developed effective improvement plans in alignment with goals and specifications.
• Created and managed project plans, timelines and budgets.