GAN based Domain Adaptation
Designed a Interactive UI based solution called PACE to feed in image and instruction to a generative model to produce a domain adapted image using our novel MAE+ GAN based model
Demo Link
Senior ML Engineer @ Bosch, India | Past - Data Scientist @ American Express
B.Tech & M.Tech (Dual), IIT Delhi 2018 batch
Bangalore, India || Github || LinkedIn
Scroll down for projects >
Designed a Interactive UI based solution called PACE to feed in image and instruction to a generative model to produce a domain adapted image using our novel MAE+ GAN based model
Demo Link
Developed a product within Bosch to retrieve historical textual requirements from given current requirements using Attention based algorithms
Demo Link
Develop a vanilla diffusion model trained on Fashion datasets to generate new cloth structures and designs
Demo LinkThis project is able to use the camera feed and generate prediction of the number shown by the human via hand sign.
Demo Link
Implements a IPM based pipeline to stich together front back and side images from a vehicle and provide a bird eye view.Additionally , we feed in camera intrinsic and extrinsic properties to improve the stitching quality.
Demo Link
Built a pipeline for consuming a YUV image stream coming from camera ( car dashboard ) and displaying its semantic segmentation on a monitor (YUV format) . This project was built on CPP environment.
Demo Link