Research & Software

Agentic workflows, AI co-scientists, and selected research and open-source projects. See also my GitHub and blog.

  • AI co-scientist

    Building agentic AI systems that act as research co-scientists: natural language interfaces for querying scientific databases, tool-calling and multi-agent orchestration (LangChain, LangGraph), and MCP integrations connecting LLM agents to data lakehouses and domain APIs—applied across genomics, microbial data, and plant–microbe research.

  • GPTgp and data lakehouse

    DOE project for foundation models in photosynthesis research: data lakehouse architecture including schema design, data modeling, and ETL for multi-modal genomic and phenotypic datasets (500+ P. trichocarpa genotypes), with MCP integrations for LLM agents.

  • AI health applications

    AI-driven health trend prediction applications with automated data collection pipelines, using React Native and Streamlit for user-facing interfaces.

  • ReLearn

    Reinforcement learning framework using Advantage Actor-Critic (A2C) for optimal control of microbial communities in bioreactors.

  • MiCoNE blog GitHub

    Scalable Nextflow pipeline for inference of microbial co-occurrence networks from 16S amplicon sequencing data, with systematic benchmarking across methods.

  • MIND

    Database with web interface and REST APIs for visualization and querying of microbial interaction networks.

  • cayenne blog GitHub

    Python package for stochastic simulations with Cython backend for high-performance computing.

  • Pipeliner GitHub

    Nextflow-based framework for bulk and single-cell RNA-seq analysis.