Explore my work in AI/ML, Bioinformatics, AI for Chemistry, and Healthcare — from autonomous research agents to deep learning and scientific computing systems.
Featured Projects
OmicsOracle AI Agent
Autonomous AI agent that automates genomic data analysis from NCBI GEO. Built with LangChain and RAG architecture, it extracts datasets, performs statistical analysis, and generates publication-ready visualizations—reducing manual analysis time from hours to minutes.
ChemAgent
Evidence-grounded pharmaceutical research assistant that connects natural-language questions to chemical and biological data sources, with verification, a REST API, and an interactive web interface.
BioPipelines
Natural-language bioinformatics workflow generator that turns analysis requests into Nextflow pipelines, with provider routing, reusable workflow components, and a Gradio interface.
All Projects
Methyl Savanna
Deep-learning methylation prediction project using EVO2 genomic foundation models, with documented data preparation, embedding, training, and GPU workflows.
Hyena-GLT
Genomic language model combining dynamic token merging with Hyena long-range convolutions for efficient DNA, RNA, and protein sequence modeling.
GenomicLightning
PyTorch Lightning framework for genomic deep learning, with modular models, sharded data handling, variant analysis, and interpretability tools.
Genomic Visualization Platform
Learning platform for production-quality genomic visualization, covering D3, genome browsers, genomic data APIs, statistical plots, and performance techniques.
QeMLflow
Machine-learning framework for chemistry and materials science, including molecular featurization, property prediction, QSAR, ADMET, and graph-based modeling workflows.
LitForge
Scientific literature search and synthesis library that unifies paper retrieval, semantic search, citation networks, ensemble scoring, and agent integrations.
LLM Evaluation Framework
Scalable evaluation framework spanning 61 models, 26 datasets, and 11 task categories, with specialized biomedical and genomics benchmarks plus GPU monitoring.
Federated Learning with Privacy
Federated-learning library and learning resource with FedAvg, FedProx, SCAFFOLD, non-IID data simulation, differential privacy, experiment tracking, and tests.
CUDA Lab
Structured CUDA programming curriculum for ML engineers and HPC practitioners, with C++-first examples, Numba alternatives, exercises, and learning tracks.
MetaFunction
Scientific paper analysis platform with DOI, PMID, arXiv, and title-based resolution, full-text extraction, AI-assisted analysis, a web interface, and a REST API.
TransMet
Deep-learning experiments for training CNN and attention-based architectures, documenting model designs for methylation-related prediction work.
GitHub Automation Suite
Automation infrastructure for genomics repositories, covering coordinated repository management, CI/CD, security scanning, health monitoring, and deployment workflows.
Clinical Text Normalization
ClinicalBERT-based encoder-decoder experiments for normalizing clinical notes, using clinical NLP datasets and preprocessing workflows for medication and adverse-event text.
Pub-Met-TL
Implementation of a transfer-learning method for completing sparse single-cell DNA methylome profiles, with training and fine-tuning workflows for methylation models.
TeamFlow
Full-stack task-management platform with multi-tenant organizations, real-time collaboration, file management, search, analytics, workflow automation, and role-based access control.
Trading ML
AI/ML trading platform combining time-series forecasting, financial sentiment analysis, technical-pattern recognition, anomaly detection, risk scoring, and API-backed dashboards.
NeuroForge
Modular neural-interface platform exploring multimodal physiological signals, brain-computer interaction, biometric authentication, medical diagnostics, and assistive communication.
Knowledge-Graph DDI Prediction
PyTorch implementation of knowledge-graph convolutional networks for predicting drug-drug interactions from drug, target, enzyme, similarity, and graph-relationship features.
GeoSpatialAI
Interactive geospatial analysis platform with progressive projects covering cartography, Census data, environmental analysis, air quality, habitat modeling, conservation, and spatial machine learning.
Data Visualization Tutorial
17-module curriculum covering visualization design, geospatial mapping, interactive dashboards, publication-quality figures, and genomic or epigenomic analytics.
PySpark Comprehensive Tutorial
Hands-on Spark curriculum spanning DataFrames, SQL, data sources, MLlib, structured streaming, graph processing, Delta Lake, and end-to-end analytics projects.
Synthea FHIR Analytics
R Shiny application for visually exploring synthetic healthcare records and extracting insights from patient, medication, and procedure data.
COVID-19 Tracker
Interactive R Markdown and HTML dashboard for presenting COVID-19 tracking and public-health data visualizations.
Genomic Variant Utilities
Python utilities for reference-genome and VCF processing, MAF extraction, clinical-variant handling, and enrichment analysis of cancer-associated variants.
FuGEP
Functional Genomics Event Prediction framework with a packaged Python implementation, configurable workflows, tutorials, tests, and deep-learning model infrastructure.
UAVarPrior
Structured variant-analysis and prioritization project with configurable workflows, notebooks, tutorials, reusable utilities, tests, and containerized development support.
AI Research Portfolio
Quarto-based portfolio and research blog bringing together AI, computational biology, genomics, publications, tutorials, and project documentation.