a final year student at IIT Jodhpur, specializing in building novel deep learning systems & products that solve complex tasks.
my passion lies in solving open problems, from AI to competitive programming.
i hate high level api wrappers :) i thrive in pressure :P- working on next generation video-to-video retrieval and visual understanding
- studied and implemented various methodologies of semantic retrieval and visual understanding
- analysed and preprocessed millions of unstructured noisy records at a stealth startup
- designed and deployed a scalable backend pipeline to de-duplicate hotel data, raising accuracy by over 15% toward industry-grade results
- built deep learning pipelines in PyTorch for cancer grading and medical imaging tasks
- contributed to multi-modal medical vision research and implemented published architectures
- implemented U-Net, LSTMs, GANs, and diffusion models from scratch
context-driven code refactoring using the Gemini API for intelligent semantic transformations. React frontend, FastAPI backend, PostgreSQL persistence.
attention-based MIL framework for cancer grading with feature fusion and generative models. trained on over 5,000 images.
from-scratch PyTorch implementation of a Transformer-based Set Prediction Network as a modular component for research pipelines.