Amit Subhash is an incoming PhD student in Neuroengineering at Indiana University, and a prolific open-source builder in the medical-imaging and AI tooling space. Across the last year he has shipped 600+ commits across 30+ public repositories, and his work spans forward and inverse solvers for Diffuse Optical Tomography (DOT), Monte Carlo photon transport (MCX, iso2mesh, NIRFASTer, NeuroDOT), domain-randomized neonatal segmentation pipelines, mechanistic interpretability of biosignal transformers, and the agentic research infrastructure that orchestrates all of it.
His central research goal is making real-time, contrast-agnostic functional brain imaging viable in the neonatal intensive care unit. He combines classical biomedical optics with modern deep learning (PyTorch, JAX), runs experiments at scale on Indiana University's Big Red 200 HPC cluster, and ships open reference implementations so other researchers can build on the work directly.
Notable open-source results include synpig, an autonomous piglet brain extraction pipeline that achieves Dice 0.9622 on 4-week piglet skull stripping, and iu-hpc-agent, a verified IU Big Red 200 / Quartz HPC reference for AI coding agents. He previously completed an MS in Data Science before starting his PhD.