Developing developer assistance AI using large language models. Previously worked on computer vision project to create single virtualized view of electricity system through aerial view imaginary.

Developed deep Learning based Intelligent Video Analytics solutions. Subtasks include,-Face Recognition-Object Detection-Scene text recognition-Automatic number-plate recognition(ANPR)which have various applications such as,-Access Control Systems-Time and Attendance System-IP Video Surveillance

Developed Deep Learning technique for accurate crop classification using very high dimensional Hyper Spectral data. My objectives included,• Deep Neural Network based Dimensionality Reduction technique which can estimate the original information with less features better than PCA.• Virtual data augmentation using statistical method for overcoming the issue of insufficient ground truth data for different crops to train deep learning model.• Comparison of MLP with CNN and SVM. Designing parallel CNNs for classifying nearly inseparable classes after separability analysis with higher accuracy.• Comparison of results from Hyper Spectral Data with Multi spectral data using model and to develop working prototype for producing crop statistics.

1. Developed a Food Recommendation API based on Collaborative Filtering machinelearning algorithm for Delicious Bunny App . Solved cold start problem for a new user by suggesting food based on friends’ reviews in app.2. Developed a Food Search API. Normalised and stored Posting List of Indian Food Names in Trie nodes for faster and accurate search. Added spelling correction using Levenshtein distance algorithm. It performs better than Swiggy’s food search which uses SQL’s LIKE keyword.Live demo: http://flash-deploy.herokuapp.com/bunny/cheese mayo

Indian Society for Technical Education students' chapter ITNU

I contributed to the development of an online food ordering app. The functionalities implemented by me are registering to a new account, dynamically loading a list of food packages available area wise, customising order, taking user review. I designed and developed the user interface from scratch.