• Bug fixes in Joplin - an open source note-taking and to-do
application.
• Contributed to express-autodocs which is
a github-action that automatically scans your codebase for express
APIs and generates a documentation website.
• Built a parser
to accept JSDoc like syntax instead of JSON comments for better
code definitions and cleaner code.
• Added as an ideas,
planning & feedback contributor to express-autodocs due to my
coding contributions and optimization suggestions.
• Received the 1st place for the NLP based News Predictor at the Enroot Mumbai
Smart City Hackathon in Oct 2018.
• As a potent tool for identifying and
curbing the false propaganda of fake news, the focus of the
project revolved around developing a web application using Django,
which would allow the users to enter any news related text or a
related image of their choice to assess the authenticity of the
information being sought.
• We made use of Tesseract
Optical Character Recognition (OCR), Gaussian Blurring and Canny
Edge detection to scan a given news article and predict its
legitimacy.
• Integrated Google Action, thereby
facilitating
Google Assistant to seamlessly interface with
our model. In addition to this, we developed a deep neural network
using bidirectional LSTMs to train our model by feeding close to
7000 articles, with appropriate labeling as real or fake.
Eventually, this enabled us to clock a 90% accuracy after
integrating multiple advanced text pre-processing mechanisms.
• Reached top 40/150 teams of the inter-college competition
‘Project Deep Blue’ in Feb 2019 for the Plastic Waste Profiling software.
•
Developed software which accepts the image of garbage clicked by a
user as input and generates a graphical analysis of the various
plastic materials in the picture and subsequently sends a
notification to the manufacturer. The object detection feature was
developed using Tensorflow and OCR.
• We presented this
project in the inter-collegiate competition ‘Project Deep Blue’,
as the use of plastic is a hotly debated issue in Mumbai and
across the world today.
• Our efforts enabled us to win a
semi-finalist spot and an abstract of our project was published in
several leading newspapers in India.