• Variational RNN with Slot Attention and pairwise updates to model the dynamics of individual objects in a scene (GRU and CNN).• Predicting the video sequences.• Technologies: Python, PyTorch, Numpy, OpenCV.

• Evaluating recent deep learning models for scene decomposition and disentanglement representation learning.• PyTorch implementation of Slot Attention for object-centric learning.• Technologies: Python, PyTorch, TensorFlow, Numpy, OpenCV.

• Automated vehicle license plate identification and recognition.• Developed spatial transformer network for the license plate rectification.• Real-time re-identification of people in difficult surroundings to extract meaningful data.• Technologies: Python, PyTorch, NumPy, pandas, OpenCV, Matplotlib, scikit-learn

• Created SAP Crystal Reports for Centricity RIS-i application catering to all European customer's requirements• Extended business use case functionalities for Oracle Reports & SAP Crystal Reports• Technologies: Oracle Databases, Crystal Reports, PL/SQL

▪ Worked on core feature enhancement of cloud-based Oracle’s property management system▪ Worked on Exchange Interfaces to provide communication between PMS and external systems, handling the end to end flow from UI creation to invoking DB functions and web services• Technologies: ADF, JSF, JSP, PL/SQL, Java, Oracle Service Bus

▪ Web service implementation (REST) to provide backward compatibility with the legacy PMS.▪ Integrated Marriott Rewards within Oracle's OPERA Cloud to be accessed across 7400 properties of Marriott International.▪ Developed 'Advance Check In' functionality for Disney which is expected to be used by 80% of all the guests.▪ Database Administrator – Maintenance

▪ Worked on Oracle’s proprietary property management system to support customization of reservation screen▪ Developed framework to facilitate the logging and debugging in the application

• Created an android application that makes the task of pathologists quite easy by automating the process of identifying whether a patient is infected with anemia.• This application extracts 26 features from the blood smear images and based on that it detects the anemic cells simultaneously classifying the type of anemia detected using the simple classifier achieving 97% accuracy.• Nominated for Gandhian Young Technological Award.• Technologies: Java, Android, Weka, OpenCV, ImageJ.

• Create an android application for a startup which is into helping communities communicate more effectively.• Technologies: Cordova, Ionic, Angular.js