Skip to content
View layaasiv's full-sized avatar

Highlights

  • Pro

Block or report layaasiv

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
layaasiv/README.md

Hi, I'm Layaa 👋

I'm a computational biologist and ML practitioner working at the intersection of machine learning and biomedical data. I build reproducible pipelines and apply deep learning and classical ML methods to problems in genomics, digital pathology, and single-cell biology.

I hold an MS in Bioinformatics and Genomics from the University of Oregon.

📫 layaasivakumar@gmail.com · LinkedIn


🧰 Skills & Tools

Languages: Python · R · Bash · SQL
Machine Learning: PyTorch · HuggingFace · scikit-learn · TensorFlow
Bioinformatics: RNA-seq · Single-cell (Seurat, Signac) · Spatial transcriptomics · Variant calling
Infrastructure: Nextflow · Docker · Git/GitHub · HPC (Slurm)


🔬 Featured Projects

Project Description Tools
Protein Localization Classifier Deep learning classifier for protein subcellular localization using ESM-2 embeddings and a custom PyTorch MLP PyTorch, HuggingFace, ESM-2
TFBS Classification & Interpretability Binding site classifiers with motif recovery analysis for CTCF and SP1 scikit-learn, LS-GKM
Digital Pathology Style Transfer Neural style transfer for H&E staining harmonization using VGG16 and CLIP evaluation PyTorch, CLIP
Single-Cell & Spatial Omics Workflows End-to-end scRNA-seq, scATAC-seq, and Visium spatial transcriptomics workflows Seurat, Signac, R
Digital Pathology Classification Fine-tuned vision models for pathology slide classification, deployed on HuggingFace Spaces PyTorch, HuggingFace
WGS Variant Discovery Pipeline Reproducible Nextflow pipeline benchmarking aligners and variant callers Nextflow, Docker, GATK

📌 Currently

🔍 Open to roles in computational biology, bioinformatics, and biomedical ML
🌱 Based in Canada · Open to opportunities in Toronto and beyond

Pinned Loading

  1. protein-location-cls protein-location-cls Public

    Protein subcellular localization classifier using ESM-2 embeddings and PyTorch MLP.

    Jupyter Notebook

  2. he-style-transfer he-style-transfer Public

    Neural style transfer for H&E staining harmonization using VGG16 and CLIP evaluation.

    Jupyter Notebook

  3. tfbs-classification tfbs-classification Public

    TFBS classification using k-mer SVM and LS-GKM with motif interpretability analysis for CTCF and SP1.

    Jupyter Notebook

  4. variant-calling variant-calling Public

    Comparing different tools used in the variant calling pipeline. Replicating the work done in this paper: Comparison of Read Mapping and Variant Calling Tools for the Analysis of Plant NGS Data (Puc…

    Nextflow

  5. digital-pathology-classification digital-pathology-classification Public

    Fine-tuned vision models for pathology image classification, deployed on HuggingFace Spaces.

    Jupyter Notebook

  6. omics-analysis omics-analysis Public

    A set of personal projects I've pursued to understand analysis of omics data.

    HTML