1. ID-FREE - Combines Biometric Virtual ID Auth for users and marketplace for Business - https://medium.com/@tnabanitade/future-of-idsin-washington-a9275081592b - Empowers Businesses to use Biometrics like Face Recognition and Fingerprint Recognition to make validate age for people as 18+ to be used as a Virtual ID for Business Interactions in Food and Restaurant Industry.

2. Masters Thesis : Detection of Fake News from Social Media.

Built Fake news detector Chrome Extension "FiB", which was featured in Washington Post, Business Insider, Mashable, Financial Times, TheNextWeb, Hacker news, The Huffington Post, the Boston Globe, Microsoft Blog, CBC News, CNBC News, CNN News, BBC News, CNET etc. - https://projectfib.azurewebsites.net 

Social Media Analytics, Natural Language Processing & Machine Learning Space

​

3. Text to Virtual Reality (Open-sourced)

This project "VR Storytellers"  was built at the World's largest VR-AR Hackathon - Virtually Reality Hackathon at the MIT Media Lab. The interface convert text (stories) in real time to Virtual Reality Experience (mapping keywords and its descriptions to objects in unity) coupled with a Background score based on the mood of the story.

Links:

a. https://www.youtube.com/watch?v=0TdY-7lcVP8&feature=emb_title

b. http://devpost.com/software/vr-story-tellers

c. GitHub Link - https://github.com/ababilinski/VR_Storytellers

​

4. Poddlebop - Connecting Mental Health Patients to Providers  

At University of Washington - Hack for Healthcare, built a Mental Health Monitoring App, which leverages natural language processing and Computer Vision to track a patient's recovery through Quantatitive and Qualitative measures.

User is asked self-assessment questions for Journal entry which is analyzed for keywords, word count, tone and writing style and data is also collected from a quick yes/no survey based on the PHQ-9 and HCL-32 peer reviewed assessment questionnaires.

This data is combined to create a metric on a scale from 0 (depressive) to 10 (manic) in order to help manage bipolar symptoms and determine treatment directives.

Won 1,25,000$ Azure Credits.

Link - https://docs.google.com/presentation/d/1Cg6ZYTuYWXdpLyLvZ5_hiP0b1eUvjZ9UxaVX2koSUC0/edit#slide=id.g35f391192_00

​

5. Building Inclusive Job Descriptions

This project was built at Microsoft OneWeek Hackathon to intuitively do automatic detection of non-inclusive text in Job Descriptions to create a much more inclusive hiring. It utilizes Natural Language Processing Techniques to tokenize and tag words and language on the fly.
 

6. Analyzing Twitter User Profiles - It involves analyzing User's personality based on user's Twitter profile given his/her twitter id, by looking at the texts & image content of User tweets to do Sentiment Analysis, Tweets Topic extraction, Named Entity recognition, political views extraction and language detection

7. Detection of Cyberbullying from Social Media using Machine Learning - Given a set of tweets, the classifier model can detect with 78% accuracy if its an instance of cyberbullying or not

8. Music Recommendor based on User's Mood from Social Networking Sites - Implemented a Music Recommendation system (Accuracy : 86%), which, given an input of User’s Twitter UserID, predicts the User’s mood as one of 20 mood classes from the User’s recent tweets using a Naïve Bayes Classifier & plays/recommends relevant mood song tracks mapped to Gaana.com. The training set of the classifier consists of 85% automatically annotated labels using mood hashtags tweets retrieval using Twitter4J & 15% manually annotated tweets and the classifier is built using 32 kinds of emoticons as well as alphabetical characters as features through Weka JAVA API.
 Compared the performance of various Classification algorithms on a smaller training dataset.
 This project’s paper won the IEEE Student Paper Contest 2015 for BPDC & is further submitted to compete in IEEE Student Paper Contest for R8 region.
 This project was also selected as top Graduation Sponsorship Program 2014 projects by Microsoft Research, ATL Cairo.

​

​

9. Text Analysis From Social Networking Sites - Implemented a system that generates trending topics from current tweets through Twitter Streaming API. Performed Sentiment Analysis on tweets, retrieved by Twitter Rest API into positive, neutral and negative classes & also rating them on a scale of 1-5. Tools used : Stanford NLP, Alchemy API

​

Chatbots

​

10. Disrupting Human Trafficking Seattle

• Worked for the NGO "Seattle Against Slavery" for their project disrupting Human Trafficking in Seattle.

• Built a Chatbot using Cortana Skills for information from the project website and further helped building their PowerBi Dashboards.

• Winners, Microsoft Hack for Good Hackathon 2018, Microsoft Garage Team

​

​

11. Cortana Money Transfer

• Built a Cortana skill to facilitate money transfer between Family using the Capital one money transfer API.

• Winners, Capital One DevExchange Hackathon 2018 and AngelHack Seattle 2018

​

​

12. LawBot 

Our team won free tickets to Techcrunch disrupt NY conference for building a chatbot called Lawbot, which using Language Modelling, Natural Language processing and Machine Learning to understand if any case has any legal implications and pertains to any law & further connects to Lawyers and Law makers like police stations, who support service on this Law.

The Techcrunch featured link is https://techcrunch.com/video/lawbot/5918915392fdde0c0d8113b1/

https://devpost.com/software/lawbot

​

13. MoodMusicBot - A Facebook Music Chatbot which maps Mood to Music

Given an Image, the chatbot can figure out the mood of the user using Microsoft Cognitive Services Emotion API and retrieves the appropriate mood based music from our Database and recommends songs.

https://www.facebook.com/getoverandletgo/?fref=ts

​

​

Slackbot Uber - Threat Response and Defense Team - A Chatbot that automates our On-Call Responses to Customers

Utilizes Natural Language Processing, to understand intents related to Security and automates responses to our Customers, just reducing oncall response time and better Customer experience.

Computer Vision, Bio-informatics & Machine Learning

14. Spoiled Food Or No.AI

Detects if a food is spoiled or not from pictures of the food. Was built as a feature to integrate with the Seeing AI App to aid the build. Built at Microsoft Oneweek Hackathon 2017.

Can be used in agriculture AI scenarios.

​

15. Fashion.AI

Detects and identifies the type of clothing, details like Ruffles and Stripes and the color of the clothing from a picture to be captioned to integrate with the Seeing AI app to aid the Blind. The Dataset was built through results from search engines on every categories. Built at Microsoft Oneweek Hackathon 2017.

​

16. Safe Pedestrian Go

A mobile app to alert users to look up while crossing the street. Utilizes the rear camera to generate Computer Vision and object detection models to identify crosswalks vs streetwalks while the user is texting and utilizes the rear camera to identify if a car is approaching, and how far the car is, while the user is on a call. Built at Oneweek Hackathon 2016.

​

17. Phoenix.AI

Built real time detection of Forest Fires and Wild Fires from Satellite Images by training an Image Classifier. Built at Microsoft Oneweek Hackathon 2017.

​

18. Road Scene Understanding for Pedestrians & Self Driving Cars through Computer Vision

Detection of Crosswalks, Pedestrians, Traffic signals, distance from other vehicles, street intersections etc from a road scene image, to use as features for the speed prediction for self-driving cars on the road.

​

​

19.  Object Relation based Image Translation using Machine Learning - Implementation of an images descriptor in Natural Language, based on the location relationships of the Objects in the Image

20. Object Recognition using Machine Learning Techniques - Implemented Supervised ML, unsupervised clustering Techniques and pattern matching through Matlab Computer Vision Lib for Object Recognition. Object classes, obtained from google, were manually labelled. Preprocessed by resizing images & extracting Speeded-Up robust features. Further Simulated by normalizing images with or without resizing

21. Implementation of Face Recognition using OpenCV - Implementation of a Real Time (WebCam video based) Multi-Face Recognition system using Eigen Vectors Method & EmuguCV (Wrapper class of OpenCV)

22. Implementation of Clustering techniques on Diabetic Retinopathy - Implemented and Compared performance (using cluster sum of square errors) of Clustering Algorithms – K Means, Fuzzy C Means and Hierarchical Clustering in clustering Diabetic Retinopathy dataset, using Matlab

23. Simulation Studies on Face Recognition using BackPropagation Neural Network - Implemented an efficient Face Recognition system by extracting features from the Datasets (through Euclidean distance formula & normalization to form preprocessed data) and then using Back propagation Neural Network technique to train the multilayer perceptron & then conducted simulation studies to minimize recognition error. Learnt Matlab & Neural Networks in the process. Used Abrosoft FaceMixer & Neuroph studio to recognize faces through neural networks & derived relation between various factors like learning rate, momentum & others with respect to Training & Testing errors by simulating values(trial & error Method.
 Won 1st Position in the interuniversity Paper Presentation Enguinity, 2014, Computer Science Category.
 Its Research paper – “Face Recognition using Backpropagation Neural Networks is accepted as one of the top 100 undergraduate research projects in UAE at UGSRC’14 - The Second United Arab Emirates Undergraduate Student Research Competition(http://www.adu.ac.ae/step-up-shabab.html).
 Research Paper accepted for Publication as well as presentation at IIENG Conference, Malaysia, 2014.
 

24. Deep Reinforcement Learning - Training Atari Games - to play against each other - Using Image Frames to train R-CNN and Q-Learning Deep Learning Algorithms to train Models to learn playing various Atari games to learn playing against a computer. 

25. Robotics : IFOR(Intelligent Flying object for Reconaissance), IFOR is UAE's first unmanned automated quadroter. I served as a Technical team member on the team - Worked on the GUI for automating functions of the quadrotor using Python, Pyside, Qt4 Designer and ROS & also on optimal Searching patterns(Exploration) for the Autonomous Quadrotor to pick up an object from a completely Unknown location within the least amount of time. 

(https://www.youtube.com/watch?v=HLdTPXp9keM

​