Developed Reinforcement Learning Model for efficient pathfinding and seamlessly integrated it
with Google Maps, enhancing real-time navigation with intelligent route optimization.
Tested BellMan Ford and Astar Model
Tech Stack: Python and Google Maps
Building Mobile application for one of the Asian Indian Festivals in Virginia providing UI to
go through the products available and option to buy the product with coupon as well as card
access to enroll into events.
Tech Stack: Flutter and TestJS
A web application using ReactJS for the VSortsTM for a Cultural Proficiency Continuum
Dialogic Protocol. Participants are taken through a set of steps to complete a survey.
Research Coordinators analyze their responses and it is sent back to the participants.
Tech Stack: ReactJS & Node.js
Designed and deployed a Programmable Logic Controller (PLC) system to monitor and secure wind energy data within a SCADA (Supervisory Control and Data Acquisition) environment. Integrated anomaly analysis techniques enabled real-time threat mitigation, protecting the renewable energy infrastructure from security vulnerabilities.
Developed a Generative Adversarial Network (GAN)-based mechanism to simulate complex attacks, aiding in tail risk assessment and providing critical insights for robust threat detection. I employed an ensemble learning-based detection scheme, achieving a 99% confidence level in defense against AI-driven attacks.
Architected Machine Learning algorithms, including Bayesian Belief Network (BBN) and Meta-Adversarial Meta-Reinforcement Learning (MAMRL) frameworks, utilizing GPU acceleration to enhance model training speeds and improve adversarial threat detection in 6G networks by 40%.
Developed a program to analyze the credibility of a candidate for loan approval based on
their previous credit history. This can eventually help the profit and the business in the
banking system
Tech Stack: Anaconda and dataset from Kaggle
Developed a web application using ReactJS and ElasticSearch.Users were able to search
abstracts for electronic thesis and dissertations and they can also upload their
paper to the site for further review and suggestions
Tech Stack: ReactJS,
ElasticSearch & Node.js
Designed a bot to have a conversation with patients in an attempt to diagnose the disease
using a small
dataset. Libraries such as Speech Recognition, win32com, NLTK, MySQL, PyAudio, Word2 number
and Winspeech
were used.
Tech Stack: Anaconda and dataset from Kaggel