Abhishek Gupta

Let's Talk is an iOS application that uses machine learning (random forests) to do early detection of language impairment in children

Key features:

1) easy to use (targeted for use by parents)
2) low-cost (smartphone based application)
3) provides detailed feedback
4) can be used for preliminary diagnosis

This application aims to solve two key problems that we see currently which is to enhance accessibility of language impairment detection in terms of cost and… Show more

Let's Talk is an iOS application that uses machine learning (random forests) to do early detection of language impairment in children

Key features:

1) easy to use (targeted for use by parents)
2) low-cost (smartphone based application)
3) provides detailed feedback
4) can be used for preliminary diagnosis

This application aims to solve two key problems that we see currently which is to enhance accessibility of language impairment detection in terms of cost and speed of intervention.

Technologies used : NLTK, language-check, Watson Speech to Text, Swoogle API, Flask, Swift, AWS, Python Show less

Medchain is a blockchain based solution to securely manage electronic health records using the authentication engine from Mobile Digital Identity (MDI). Specifically, medchain enabled healthcare attributes consent management on the blockchain using smart contracts. Built at the Ericsson Spring Hackathon 2017.

Technologies used :

Monax, Solidity, Express, NodeJS, Flask, Python, Twilio

The Thinking Cap helps you to supercharge your learning. It is a truly adaptive learning analysis system for accelerated and extended flow states. It leverages neural networks to train on personalized intricacies of your EEG and helps you learn better habits so that you can create your Pieta. The Thinking Cap is grounded strongly in neuroscience, extending the benefits of academic grade research to you in a fun and interactive user experience.

Traveling to a new city always comes with a million choices on a million websites that need to be looked at to make the best of your experience in that city. We want to change that and allow you to show up to a new city, open REGO and then select "What you are in the mood for?" and Voila! You have an itinerary for the entire day giving you activities and meals. Say NO to research and YES to serendipity.

A technically sophisticated solution, REGO is built on a Flask server running on EC2… Show more

Traveling to a new city always comes with a million choices on a million websites that need to be looked at to make the best of your experience in that city. We want to change that and allow you to show up to a new city, open REGO and then select "What you are in the mood for?" and Voila! You have an itinerary for the entire day giving you activities and meals. Say NO to research and YES to serendipity.

A technically sophisticated solution, REGO is built on a Flask server running on EC2 with a pipeline to process data as follows:

- resolving universal city name
- local caching of city data
- categorization of Expedia data
- keyword analysis
- mood to keyword mapping based on semantic distance and 'bucketing'
- Yelp queries limited by radius in proximity to the activity
- activity selection based on the mood and then the final itinerary output Show less

Sales teams generate hundreds of gigabytes of data a year as the profession becomes increasing mobile. Rather than let it go to waste, we thought chat transcripts along with their "reactions" could be used to create a real time map of engagement, morale and feedback. By pairing these qualitative findings with the corresponding sales data, Sentient gives sales managers the full story: here's how sales are doing, and here's how the team feels. Human Ressource Managers can follow trends over time,… Show more

Sales teams generate hundreds of gigabytes of data a year as the profession becomes increasing mobile. Rather than let it go to waste, we thought chat transcripts along with their "reactions" could be used to create a real time map of engagement, morale and feedback. By pairing these qualitative findings with the corresponding sales data, Sentient gives sales managers the full story: here's how sales are doing, and here's how the team feels. Human Ressource Managers can follow trends over time, search by key word and zoom into individual users.

Built using Python, Flask, EC2, machine learning, keyword and sentiment analysis Show less

Geographic analysis of vegetation distribution along with soil contamination level assessment to recommend which species to plant where for Ville de Montreal.

Winner at HackCite hosted by Desjardins Labs

We estimated the influence of word-of-mouth on businesses as a function of the sum total of their reviewers' word-of-mouth potential for that business. We based users' word-of-mouth power on a simple criterion: how many of a user's friends visit an establishment after the user gives it a positive review? This varies from establishment to establishment because different types and locations of business are subject to different pressure from word of mouth.

Built using Python, EC2, Yelp… Show more

We estimated the influence of word-of-mouth on businesses as a function of the sum total of their reviewers' word-of-mouth potential for that business. We based users' word-of-mouth power on a simple criterion: how many of a user's friends visit an establishment after the user gives it a positive review? This varies from establishment to establishment because different types and locations of business are subject to different pressure from word of mouth.

Built using Python, EC2, Yelp academic dataset, Google Places API,machine learning, keyword and sentiment analysis. Show less

Creating a marketplace using Instagram allowing brands to search for potential advertisers based on ‘social influence’ score calculated by mathematical model factoring in geo-location; built using Indico ML API, Flask, iOS.

An indoor-location based shopping solution allowing personalized offers and checkout at shelf; used Flask, Estimotes, iOS and proprietary Ericsson security engine

A seamless parking mechanism using LoA3(level of assurance) with drive-in and auto payment ability; built using Flask, NodeJS, Android and proprietary Ericsson security engine

An iOS & Flask app to help people in Montreal explore culture/heritage/events through random discovery/trending/mood options

A 2-factor authentication payment mechanism using heartbeat-based authentication and MasterCard APIs; won NymiBand for use of their wearable device; used Python and iOS

- Developed at AngelHack Montreal 2015 (Winner)
- An application for facilitators in schools and managers in office to boost productivity by analyzing stress signals
- Used an SVM classifier to categorize brain signals from the Muse Headband to allow display on the dashboard
- Created abstraction layer to work with front-end component to interface with any Brain Computer Interface (BCI) device

- Project at the Expedia Montreal Hackathon (3rd Place Winner)
- Flaneur is a travel-companion app that will suggest daily itineraries based on your mood.
- Developed the interface for the app on Cordova/Phonegap using the App.js library
- Use of geolocation data which was sent to the backend for analysis
- Backend requests processed using Flask (Python framework)

- Creating a Pokedex (database of Pokemon) for the new Pokemon game
- Used Ruby to scrape website for information on the Pokemon
- Android SDK to implement the app for smartphones

• Create an Android App in Java/Android SDK which allows users to share photos in a feed and show off exciting locations they've visited using photo geolocation data
• Design a sleek, simple and easy-to-use user interface in XML
• Develop the mechanism which opens the phone's camera, saves a photo, and allows it to be shared on the feed
• Implement the interaction between the phone and the server, which allows the user to upload their photos or browse the feed

-- 5 Person Team developing a strategy tile based video game application
-- Client/Server model to support multiple players over a network
-- Created a layered architecture for scalability and extensibility
-- Created a custom serializer/deserialzer to communicate game objects over the network