Working on ML based solution to remove tint (uneven color patches) from the image created due to uneven roll-off gain correction.

Lead all the technical developments for the camera simulator team for Snapdragon 8 Gen 3.

▪ Leading developmental efforts for the company’s Facial Recognition System and Facial Recognition based Attendance Management System.▪ Lead engineer for development of Intelligent Video Analytics Applications for Smart Cities.▪ Design and development of Awiros Spira- a platform for helping developers in quickly and easily creating Intelligent Video Analytics applications.▪ Lead engineer for design and development of Awiros Training Orchestration Module- A platform for annotation, training and subsequent deployment of Deep Learning models

▪ Explored methods for estimating homography of single dominating plane in image.▪ Developed new dataset consist of almost 8k building images using Google StreetView from which wewere able to create nearly 45k pairs of images.▪ Developed CNN architecture from scratch on pytorch to train the model on created dataset.▪ Performed various experiments to analyze the performance of our model.

▪ Improved their existing optic disc classifier in retinal fundus images by 20%▪ Developed a new system for finding optic disc contour using online regression technique which was useful for screening glaucoma (blindness) cases.▪ Created a classifier for classifying whether a retinal image is of good or bad quality.