M.S. in Computer Engineering (Colorado State University, expected Spring 2026) — Thesis: Lightweight Topology-Aware Graph Learning for Green AI and Edge Deployment.
Embedded systems + ML, with an emphasis on efficient edge deployment
I work at the intersection of hardware/embedded systems and machine learning, with a strong interest in Green AI—reducing parameters, memory footprint, and latency while keeping models practical for resource-constrained devices (e.g., Raspberry Pi-class deployments).
(Customize this list further with the exact frameworks you use: PyTorch/TensorFlow, RTOS, FPGA toolchains, etc.)
Selected work from embedded systems, assistive robotics, and applied ML
Built a GNN-free graph learning pipeline using feature-engineered embeddings, autoencoder compression, and MLP + knowledge distillation for edge-efficient node classification, targeting reduced parameters, memory, and latency on OGBN-Arxiv/Proteins.
Enhanced a robotic assistive frame for paralyzed dogs. Identified/modified motors to reach required torque for hip and knee actuation, integrating control through a Raspberry Pi 4 for precise motion coordination.
Built a medical image processing pipeline for non-invasive diabetes detection using iris segmentation, feature extraction, and ML classifiers (ANN, SVM) for early diagnostic support.
Raspberry Pi–based system integrating speech recognition and NLP to classify users into houses from voice responses. Performed audio feature extraction and optimized the model for low-latency, resource-efficient edge execution.
Academic leadership and applied engineering experience
Selected publications and works-in-progress
Graph feature engineering and embeddings for AIoT, with an emphasis on efficient learning pipelines suitable for energy-aware and resource-constrained deployment.
Service, committees, and professional affiliations
-
Interested in internships, research collaborations, or embedded/edge ML engineering work. The fastest way to reach me is email.
If you want, add a button here to download your resume PDF (place the PDF in the same folder as this HTML and name it Yash_CSU.pdf).