15 million American's have the Alzheimer's disease. Our cross-platform app is a first-of-its-kind personalized assistant for such individuals, allowing them to maintain their memory in the short and long-term with additional features for safety.

One of the key features of Alz is a native "visual contact base" built with the Microsoft Azure Face API to allow for the facial identification of immediate family members of the user, a task which may prove to be diffcult at times for such indivduals. Other features include a map to guide users back to the safety of their home.

The front-end is completely built in Flutter, with a focus on making an accessible design that can be used by all ages. Our backend is completely asynchronous, written in python with the aiohttp REST framework, availing the Microsoft Azure cognitive services, and hosted on Heroku cloud.

Check it out on Github!

There is a lot of information out there and simply, not enough time in our daily lives to go through all the video publications that are available. To overcome this problem, Recap This Video was developed.

This project embeds a web app, with a serverless backend which transcribes the videos and produces a summary for the user using NLP. To produce a list of top sentences to be included in the summary, the NLP algorithm uses a similarity matrix and graph visualisation, with the pager rank algorithm. Key libraries used: NLTK, numpy, networkx.

To pre-process the video transcriptions for inaccuracies, the python Natural Language ToolKit (NLTK) was used. A similarilty matrix was then developed using cosine similarity. A graph was then constructed using networkx, with the nodes as the sentences and the edges as the similarity scores. To arrive at the top sentences, the pager-rank algoritm was used.

Check it out on Github!

Developing an extractive summarization model with TensorFlow, and deployed on the browser with TensorFlowJS with low-latency predictions. This is a novel idea where users can take a screenshot of their screen and the extension will automatically summarize the contents of the screenshot.

A cross-platform web app with geo-locating features and custom data analysis, to provide a safe platform for trick or treaters to interact with houses (in their neighborhood) handing out candy even during the pandemic.

The data analysis was powered with a Microsoft Covid dashboard indicating levels of safety at a county, state, and country level, and then scraped with Python to provide real-time results for users. The front-end was developed using vanilla JS, HTML/CSS with a backend powered by NodeJS Express.

Check it out on Github!

Developing an open-source social media platform to connect university students on co-op, offering unique features like a housing marketplace, in-app messaging with merchants, event posts, surveys, and communities.

Check us out on Github!

An NLP-based web app which can be used by HR departments, and the police, to generate a character analysis based on a person's social media handle.

The Web App was built on a React front-end, a python flask back-end, and MongoDB. To obtain the social media data, the twitter API was used to check for recent and popular tweets. To generate the character analysis, the google cloud natural language processing API was used, classifying user emotion associated with recurring themes, locations, and words. The social media data and the results were synced with a person's profile on MongoDB, expediting the analysis process in future steps.

Check it out on Github!

This project presents a programmatic solution to the knight's tour problem, an age-old mathematical puzzle which seeks to find if a knight can traverse every square on a chessboard (of that size) exactly once. The solution to this problem was developed using a Java algorithm.

Check it out on Github!