Ajay T AI / Computer Vision Engineer

I build clean, reproducible implementations of deep learning models for computer vision — from CNNs to Vision Transformers — and document the journey.

View Projects DL for CV Repo Download Résumé

Socials

Featured Projects

Vision Transformer (ViT) from Scratch

PyTorch

Patch embeddings, multi-head self-attention, and training on CIFAR-10 with strong augmentation.

transformers cifar10 timm-like

View code →

UNet for Medical Segmentation

PyTorch

Encoder-decoder with skip connections, Dice loss, and mixed-precision training.

segmentation dice-loss amp

View code →

YOLO-style Detector (Tiny)

PyTorch

Anchor boxes, CIoU loss, NMS pipeline, and training