Johns Hopkins University · M.S.E. in ECE
A software engineer and ML researcher focused on deep learning, computer vision, and multimodal AI. From fine-tuning LLMs on medical data to deploying real-time models on autonomous vehicles — I bring research ideas into production.
Chapter One
Chapter Two
Implemented ResNet (2.8M params) achieving 93% test accuracy; trained ViT with cosine LR scheduling, mixed precision, and systematic convergence analysis on large-scale image data.
Built PyTorch → ONNX → TensorRT deployment pipeline; applied INT8 post-training quantization achieving 4.17× throughput improvement with only 0.6 mAP drop and 70% model size reduction.
Implemented and evaluated 6 classical algorithms (MAP-L2, ROF, TGV2, Non-Local Means, Bilateral Filter) across 50 COCO images with statistical performance analysis.
Trained Nerfacto model on a 59-view custom dataset; compared volumetric NeRF vs. Gaussian Splatting on rendering efficiency, memory footprint, and reconstruction fidelity.
Chapter Three
Chapter Four
Foundation