Hello, I’m Soham.

About

I’m a junior studying Biomedical Engineering at the University of South Florida, and am interested in working at the intersection of machine learning and neuroscience.

My primary focus is connectomics and computational modeling, where I aim to build tools to understand how neural circuits represent, transform, and store information. I’m especially interested in bridging biological and artificial systems by developing methods that can allow us to ask questions we can’t think of yet.

More broadly, I’m curious about problems where computation meets biology, including bioinformatics, computational drug discovery, functional imaging, and neural population analysis.

Outside the lab, I work as a Resident Assistant and co-founded a student-led STEM research community that helps students find collaborators, start projects, and explore niche topics! I’m also a part of the bioinformatics club at USF and enjoy running, reading, working out, and falling from my skateboard!

Work

Research Intern

Auditory Development and Connectomics Lab

Co-developed an automated neuron segmentation pipeline, turning large EM volumes into interactive, analysis-ready data and drastically reducing manual tracing effort.

Resident Assistant

University of South Florida

Mentored students and built a supportive community through academic and wellness programming, while serving as a first point of contact for student concerns.

Research Trainee

Moffitt Cancer Center

Improved cell segmentation model performance by changing training strategies and parameters, and built a structured data pipeline.

Publications & Projects

Machine Learning and Molecular Docking Prediction of Potential Inhibitors against Dengue Virus

Contributed to model development using classical machine learning methods to analyze biological datasets and compare predictive performance across approaches. Explored molecular dynamics and docking workflows to study molecular interactions, and assisted in preparing results and figures for manuscript completion.

JupyterLab, scikit-learn models, PaDEL descriptors, AutoDock Vina, Molecular dynamics (MD) simulations, PyMOL.

BrianRoute - An Open Machine Learning Platform for Blood–Brain Barrier
Permeability Prediction

Built and deployed a web platform for predicting blood-brain barrier permeability of small molecules, allowing users to submit compounds, retrieve predictions, and explore a growing dataset of results. Led an international team to design the full-stack system.

Machine learning (KNN, RF, Logistic Regression, XGBoost, LGBM), Streamlit, LLM API integration, React-JS, Google API integration

NCBI and other Bioinformatics Codeathons

I have also participated in multiple NCBI and Carnegie Mellon Libraries codeathons, some of which have led to published preprints. Please check out my LinkedIn page, where I’ve described these projects and my contributions!

Check out my complete research resume here

Resume

Let’s connect

If you’re interested in collaborating - or just want to chat about neuroAI or bioinformatics - feel free to reach out !!

Email LinkedIn GitHub Twitter

Note

The fastest way to reach me is through email. I’m currently working on this website and plan to add a lot. Stay tuned!