Formulate a multi-agent single human teaming problem with codependency to achieve a common goal with the aim of increasing the team utility. Developed and tested suitable use case to capture teaming aspects of human with AI agents.

Developed a fully functional interactive dashboard for state administration using Flask framework.

Department and Rank System: Departmental heirarchy with Covid-19, Judiciary, Police and Schemes dashboards, with rank based user access to different parts of Dasboard.

Real-time graphing and Analysis: Time Interval Filtering, region filtering and many other data-specific options to analyse data pictorially.

Trained LSTM on MIDI files and further used Q-, Psi- and G- learning to fine tune the learned RNN parameters using varying reward functions for two emotions- happy and sad. The rewards were adapted from Music Theory.

Attendance system, which automatically identifies driver information using face recognition and number plate using number plate detection. Developed using Django backend and React.js frontend.

Used an InceptionV3-LSTM Encoder-Decoder system with Bahdanau Attention to caption images clicked by the Blind in their everyday lives, thus, often being blurred and unclear. Implemented several techniques for Captioning with this performing best.

Implemented IDA* Search with obstacles on a grid between a start and end point. Also worked on developing the interface of the web-app.

Developed a portal for Railway Administration with Database optimisation. The portal allows addition of same train with different number and types of coaches on different dates and allows for multi-agent registration to book tickets and generate pnr, if seats available.

Developed pipelined RISCV-ISA simulator for parallel instruction execution with dataforwarding, bit prediction and inter-state caches.

Developed an emotion recognition and classifier system on non-posed images of people on 6 emotions.

Given news pieces and their headlines, developed an LSTM based news classifier, with vocabulary derived from WikiText-103 to categorise the news pieces under different subject headings.

A blogging website with user signup, login functionality, built on django. Project to experiment with the django platform.