Working on extreme weather events prediction.

Working on applying AI for biodiversity conservation and climate science with Prof. David Rolnick.

Worked on emulating radiative transfer using deep learning models using physics-constraints.

• Working on the problem of change detection in geospatial imagery using deep learning and computer vision techniques.• The input is a pair of images of a location taken at different timestamps. Theoutput is the change between the two images.• Currently trying out different approaches to improve the entire pipeline andbuild a proof of concept for a few potential clients.

• Researching primarily in the field of artificial intelligence and computer vision to apply cutting-edge technology to real-world scenarios.• Working on multiple ongoing research projects.• Currently, the lab has advisors from TensorFlow, Ubisoft, MIT and University ofPittsburgh.

• Worked on pilferage detection using CCTV footage provided by the client. The data was of low quality and had to be improved using image processing methods like sharpening and adaptive histogram equalization.• Made use of Faster R-CNN architecture with TensorFlow object detection API and used various architectures feature extractors like Inception and Inception-Resnet. Tried out various optimizers like Adam and SGD along with Data Augmentation techniques for increasing the trainset size… Show more • Worked on pilferage detection using CCTV footage provided by the client. The data was of low quality and had to be improved using image processing methods like sharpening and adaptive histogram equalization.• Made use of Faster R-CNN architecture with TensorFlow object detection API and used various architectures feature extractors like Inception and Inception-Resnet. Tried out various optimizers like Adam and SGD along with Data Augmentation techniques for increasing the trainset size using Imgaug library.• Model achieved a mAP (Mean average precision) of 0.887.• Major libraries used for the project were TensorFlow, OpenCV and Imgaug. Show less

• Developed a machine learning model for the project 'Employee Appraisal'.• Demonstrated the use of supervised machine learning techniques to predict which employees are worthy of a promotion.• Achieved an accuracy of 85% on a validation set.• Major libraries used for this project were Scikit-Learn, Pandas, Numpy and Matplotlib.