- I'm a masters student in Biomedical Informatics at Stony Brook University
- Interested in large language models and their application in bioinformatics
- My capstone project is using DNABERT-S to build better metagenomics tools and learn more about the human microbiome
- I'm currently looking for an internship in bioinformatics or machine learning engineering for summer 2025
- Here I have various individual and university projects that I've worked on throughout the years
- DNABERT-S Weighted Sum Pooling Implementation is a project that implements a learned global attention network to generate weights for each token-level embedding before pooling together to understand which parts of the sequence are more meaningful for species discrimination.
- COMEBERT is a combination of the metagenomics binning program, COMEBin, and the pretrained DNA language models, DNABERT (1, 2, and S). My modification replaces tetra-nucleotide frequency embeddings with different DNABERT embeddings, and comebines them with sequencing depth information to bin sequences in an unsupervised manner.
- EmerGNN model was a graduate project that allowed me to learn more about graph neural networks and web development
- Lineup reader website was an individual project to self-learn web development using Flask
- My Neural Network Implementation was an individual project to learn about OOP, and various deep learning optimizers, loss functions, and preprocessing techniques
- Lysis-lysogeny gene model was an undergraduate project to model gene expression of bacteriophages and helped me learn about statistical methods like Monte-Carlo simulations