I'm Joshua Pickard, a Eric and Wendy Schmidt Center postdoc at the Broad Institute of Harvard and MIT. I work with the Center for Integrated Solutions to Infectious Disease (CISID) and collaborators across Mass General Brigham and Harvard to develop control theory methods and AI-based digital twins of sepsis and crticial care patients.

Previously, I trained with Indika Rajapakse at the University of Michigan, studying bioinformatics and digital biology for cellular reprogramming. I am passionate about investigating dynamic biomedical systems across a wide range of data and scales 🚀.

Below you can find updates on my work and research areas.

• July `26 Our paper EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins has been accepted to Machine Learning for Healthcare (MLHC) 2026.
• May `26 — Chalk talk titled Real-time Digital Twins for Sepsis Control at the Broad's Models Inference and Algorithms (MIA) seminar series
• May `26 — Flash talk titled Real-time Digital Twins for Sepsis Control at Harvard EECS Research Day
• Apr. `26 — Our paper Data-Driven Tensor Decomposition Identification of Homogeneous Polynomial Dynamical Systems is available in preprint.
• Mar. `26 — Our paper Structural Controllability of Large-Scale Hypergraphs is available in preprint on arXiv.
• Jan. `26 — Organized the AI4ID Symposium and led a primer session on AI architectures.
• Dec. `25 — Our paper HIF: The Hypergraph Interchange Format for Higher-Order Networks is available online in Network Science.
• Oct. `25 — Our paper Dynamic Sensor Selection for Biomarker Discovery is available online in PNAS
• Oct. `25 — Nominated for the U-M ProQuest Distinguished Dissertation Award by the Bioinformatics Graduate Program.
• Sep. `25 — Started as a Postdoctoral Fellow at the Eric and Wendy Schmidt Center at the Broad Institute.
• June `25 — Defended my PhD thesis at the University of Michigan.
• May `25 — Our paper Automatic biomarker discovery and enrichment with BRAD is available online in Bioinformatics
• Mar. `25 — Nominated for the U-M Medical School Excellence in Teaching Award by the Bioinformatics Graduate Program.
• Sep. `24 — Our paper Kronecker Product of Tensors and Hypergraphs: Structure and Dynamics is available online in SIAM Journal of Matrix Analysis and Application (SIMAX)
• Aug. `24 — Our paper Continuous sepsis trajectory prediction using tensor-reduced physiological signals is available online in Scientific Reports
• June `23 — Our paper HAT: Hypergraph Analysis Toolbox* is available online in PLOS Computational Biology
• Sep. `22 — Our paper Deciphering multi-way interactions in the human genome is available online in Nature Communications

• 🤖LLM Powered Digital Biology: AI isn’t just changing what we know but how we work. My Bioinformatics Retrieval Augmented Digital assistant (BRAD) agentic chatbot leverages language models to automate workflows like information retrieval (online), software execution (local), and document search (RAG) to accelerate digital biology research.
• 🔍Biomarker Observability: With recent and rapic advancements in experimental methods, data processing has become a bigger challenge than acquisition. I’m working on observability-based methods to identify meaningful biomarkers in genomic data 🧬.
• 🌐Higher Order Networks: Biological data often needs unconventional analysis techniques. I develop methods and tools like the Hypergraph Analysis Toolbox to study group interactions within genomic networks and beyond.

Thanks for stopping by my profile! Feel free to reach out (📫jpic@umich.edu) if you’re interested in discussing research or have a cool idea to share.