I build real-time systems, scientific data pipelines, and ML tools for neuroscience research.
Currently a neuroscience researcher at the University of Michigan Ann Arbor, with two years embedded in labs building closed-loop neural systems, fMRI infrastructure, and neural decoding workflows.
Real-time neural systems
Closed-loop VR for head-fixed rodents, sub-16 ms feedback, <1 ms timing jitter, and hardware
synchronization across electrophysiology, behavior tracking, and stimulus delivery.
Scientific data pipelines
Reproducible fMRI preprocessing workflows, DVC-backed versioning, CI/CD, and scalable
neuroimaging infrastructure for cross-sectional and longitudinal cohorts.
ML for neural data
CNN-based brain-state decoding, Grad-CAM model interpretation, SLEAP pose estimation, and
kinematic extraction across 100+ behavioral sessions.
| Project | Focus | Tech / Results |
|---|---|---|
| SLEAP Pipeline Manager | Desktop workflow manager for SLEAP labeling, Great Lakes HPC training, inference, and review | Python, Tkinter, SSH/SFTP, Slurm, PyInstaller, interactive Duo auth, task history, configurable inference profiles |
| RythMice | Behavioral analysis toolkit for rodent locomotion rhythm studies | Python, MATLAB, LabJack-compatible traces, CLI + GUI, QC plots, session plans |
| FMRI-image-ML-classifier | Visual stimulus decoder from fMRI cortical activation maps | PyTorch, ResNet-18, MobileNet, subject-level splits, 94% F1, 0.97 ROC-AUC |
| Resume-Builder | Browser-based resume builder with PDF/DOCX export | Next.js, React, TypeScript, local-first storage, live site at jobcraftkit.com |
Python PyTorch MATLAB C# / Unity TypeScript React Next.js
Docker GitHub Actions DVC AWS Linux
fMRI SPM CONN SLEAP Biopac LabJack
- Real-time closed-loop systems for neuroscience experiments
- Reproducible neuroimaging and behavioral data pipelines
- Interpretable ML for neural and experimental data