Developed a generative AI based approach for video editing to automatically replace protagonists in advertisement videos using text prompts, reducing production costs for creating video marketing content.

Worked with the Palette AI Research (PAIR) team at Amazon Advertising on building machine learning models that can automatically create high-quality and high-performance video advertisements.

Worked with Prof. Liam Paull on continual learning for neural coordinate maps such as NeRF. Also explored self-supervised depth estimation leveraging the gradSLAM framework.

Worked with the Fusion Analytics Warehouse (FAW) team on developing a framework for extracting and analysing customer usage data to offer actionable insights and drive informed product development decisions.

Worked at the Video Analytics Lab with Prof. Venkatesh Babu and Prof. Varun Jampani (Google Research) on self-supervised keypoint detection from category specific image collections [work accepted at WACV 2022]. Also, worked on approaches for self-supervised single-view 3D reconstruction.

Worked with Prof. Deepak Mishra, IIST, Trivandrum on second-order pooling mechanisms for graph neural networks.

Developed robust models for segmentation of Focal Cortical Dysplasia (FCD) lesions from Magnetic Resonance (MR) images with Dr. Jeny Rajan. Also explored deep learning based methods for detecting defective PCBs.

Tejan Karmali*, Abhinav Atrishi*, Sai Sree Harsha, Susmit Agrawal, Varun Jampani, Venkatesh Babu