Shraddha Piparia, Ph.D.

Applied Scientist & Computational Biologist

Modeling disease heterogeneity with machine learning.

I build machine learning models across diverse biomedical data types including genomics, miRNA, proteomics, and clinical data to characterize study treatment response.

My work focuses on patient stratification, drug response modeling, and representation learning for high-dimensional biological data.

GitHub LinkedIn

Experience

Postdoctoral Scientist, University of California San Diego, USA (Nov 2021 – Present)

Computational genomics and pharmacogenomics research focused on disease heterogeneity, patient stratification, and treatment response modeling in asthma and immune-mediated disease.

Developed LD-aware representation learning models for high-dimensional genotype data to improve patient stratification and drug response prediction.
Identified neurocognitive subtype–specific protein signatures using Olink NPX data and scalable UK Biobank pipelines.
Built ML pipelines for large-scale genomic, transcriptomic, and clinical datasets.
Developed VAE-based genotype embeddings for patient stratification and pharmacogenomics.
Modeled asthma heterogeneity and ICS treatment response using multi-modal data.
Designed scalable workflows (PLINK, R, Python, HPC) for reproducible genomics analysis.

Application Developer, Oracle India Private Limited®, Telangana, India (Jun 2013 – Jul 2016)

Software engineer working on Oracle Cloud Enterprise applications (Oracle Project Portfolio Management).

Designed and implemented production features for Oracle Cloud enterprise applications.
Developed Build Verification Tests (BVTs) to improve release reliability and software quality.
Built a machine learning–based sentiment analysis tool using enterprise communication data (Oracle Social Network) to improve search relevance and project outcome prediction.

Education

Ph.D., Computer Science, University of North Texas, USA

Focus: Machine learning and probabilistic modeling for discovering structure in complex, event-driven data (CRFs, neural sequence models).