Projects

🔬 Projects

fNIRS-DataProcessing Pipelines

Modular fNIRS preprocessing & analysis pipelines for language and pain studies.

• Tech stack: MATLAB (NIRS Toolbox)
• Highlights: Applied to datasets of 600+ participants; includes preprocessing utilities, GLM-ready design matrices, first and second level analysis, results plotting, and QC helpers

fNIRS + NLP

Exploring brain–language alignment using fNIRS recordings and large language models (LLM).

• Tech stack: MATLAB (NIRS Toolbox), Python (PyTorch, Hugging Face Transformers)
• Highlights: Compares children’s fNIRS responses during story listening with GPT-2 model embeddings or generated surprisals; bridges cognitive neuroscience and AI
  

PainTrek / DocTrek (Private code)

AI-assisted pain tracking and documentation platform.

• Tech stack: Python (FastAPI), LM Studio, Faiss, Docker, LLM
• Highlights: Patient–clinician summaries, HIPAA-compliant deployment, integration with EHR systems