Engineer and product thinker building at the intersection of wearables BCI, ML algorithms, and AI products. Currently finishing my M.S. in Biomedical Engineering at Carnegie Mellon, with a B.Eng. in Computer Engineering from McGill.
I've led cross-functional teams shipping wearable devices (Parkinson's tremor exoskeleton, combat triage vest, PT rehab sleeve), built the ML pipelines behind them (BCI decoders, sensor fusion, signal processing), and owned the system architecture connecting sensor data to real product decisions. Currently conducting research at CMU's Neuro-Mechatronics Lab.
| Domain | Project | Description | Tech Stack | Link |
|---|---|---|---|---|
| π§ Neuro/BCI | SMR BCI Decoder Optimization | Offline optimization of sensorimotor rhythm BCI decoders using CSP+LDA and EEGNet. Achieved 90.9% motor imagery classification from 64-channel EEG | View | |
| π¦Ύ Wearables | Parkinson's Exoskeleton | Wearable exoskeleton with ML-based tremor detection and suppression. Led 5-person team as Team Captain | View | |
| π¦Ύ Wearables | PT Rehab Sleeve | Physical therapy sleeve for post-injury/stroke muscle asymmetry correction with AI-generated exercise suggestions and progress analysis | View | |
| π¦Ύ Wearables | IMU Data Harmonization | Data harmonization pipeline for IMU biomechanical sensor data across heterogeneous devices and protocols | View | |
| π¦Ύ Wearables π€ AI/ML | Combat Injury Triage | Wearable sensing vest with hybrid ML triage scoring system for real-time combat casualty assessment | View | |
| π€ AI/ML | TinyLlama Optimization | LLM optimization via PyTorch quantization, mixed-precision, and FPGA-accelerated SoftMax. 8Γ model size reduction, 20% inference speedup | View | |
| π€ AI/ML | Semantic Image Segmentation | U-Net and DeepLabV3 segmentation on Oxford-IIIT Pets. 12% improvement in Dice coefficient over baselines | View | |
| ποΈ Computer Vision | Feature Matching & Stitching | Harris Corner Detection, SIFT feature matching, and RANSAC-based panoramic stitching | View | |
| ποΈ Computer Vision | Classifiers & Object Detection | MNIST digit classification (SVM/Random Forest) and YOLOv5 street scene detection | View |

