Research
PiCME: Pipeline for Contrastive Modality Evaluation
Singh Lab at Brown University
Developing a multi-modal contrastive learning framework for integrating health data in the MIMIC-IV dataset. Collaboration with Michal Golovanevsky.
SRO: Subspace Relaxation Operator for Protein Structure Prediction
with Colin Baker and Sorin Istrail
Creating a trainable operator that uses solvent-specific forces to guide protein structure prediction towards favorable configurations.
BindGPS: Investigating 3D Chromatin Formation
Larschan Lab at Brown University
Designing a multiomic dataset with Micro-C, ChIP-seq, and sequence data for analysis using GNNs and LLMs.
BindCompare
Larschan Lab at Brown University
BindCompare is a novel integrated protein–nucleic acid binding analysis platform. It is a user-friendly tool that can be run locally to predict new combinatorial relationships between TFs and RBPs.
DNABERT-Enhancer
Ovcharenko Group at the NLM (NIH)
Developed a transformer-based model for enhancer prediction by fine-tuning DNABERT on enhancer sequences. Implemented novel attention visualization techniques to improve model interpretability and biological insight.
Wet Lab Research
Larschan Lab at Brown University
Collaborated on multiple projects investigating genome-wide sex-specific splicing during early Drosophila embryogenesis and the role of dual DNA/RNA-binding factors in regulating hnRNP splicing condensates. Combined wet lab experiments with computational analysis to reveal insights into developmental biology and splicing regulation.