ContApp is an Android app that allows for a new way to visualize contacts. Traditional contact apps show you the list of people you know in a boring, non-intelligent, alphabetical fashion. But this app, makes use of the latest and greatest that the Android SDK has to offer and provides the user with a file-and-folder (heirarchy-based) view to easily view and manage their contacts! And the things this app can do once you have such structured data are amazing, to say the least 😏
Android (frontend): Android Studio Kotlin MVVM Room Retrofit RecyclerTreeView (3rd party lib)
Backend: Node.JS MongoDB Atlas Express JWT🖼 Open Gallery Source (still private)
I firmly believe that education is my passport to the world, and so I embrace continuing education - it is one of my biggest lifelong ongoing projects. You can take a scroll through the following slides with all of my LinkedIn Learning certifications, or view them on my LinkedIn profile.
Tool(s) used: LinkedIn Learning
As a part of my Entrepreneurship minor, I took up the course "Entrepreneurship: Principles and Practices (ECON 325)" at UNC, through which, I have mastered the art of doing quick, back-of-the-envelope cash-flow statements. As can be observed through my LinkedIn learning certifications above, I have also done a course that goes into the details of Income, Profit/Loss, and Cash flow statements. Here are some of the cashflow statements that I have made in order to demonstrate my skills:
Tool(s) used: Spreadsheets
As a part of my Principles of Entrepreneurship course, I learned and applied rapid prototyping techniques to come up with two 3D designs in Tinkercad. I created first a rough design of a COVID-19-aware Kiosk system, which can perform face and finger recognition, can print tokens to act as a queue manager, and, last but not least, has a sanitizer dispenser (the magenta bottle to the left). Then, I created a rudimentary design for a robot that can perform surveillance via a camera. I believe it can become a great product for facilities where security is essential, such as prisons.
Tool(s) used: TinkercadView Kiosk on Tinkercad View Robot on Tinkercad
Through the course Data Science in the Business World, offered in Spring 2021 jointly by UNC's Kenan-Flagler Business School and the Department of Computer Science, I have come understand the prominence of unstructured data in the digital world and the potential of Natural Language Processing. Here is an infographic that explains my thoughts in further detail.
I wrote my first resume when I was applying for a university chapter membership back in November 2017. It was shabby and other than my academic achievements and small robotics projects, I did not have much to show. I have worked hard on developing my skillset and establishing myself as a skilled software developer who takes pride in his work, and is constantly on the lookout for learning new things and growing. The following resume is a result of those years of sweat and perseverance.
Samosabucket is an open-source (GPL v3) food delivery web app which is easy (and free) to set up, and handles the full workflow needed to run a basic food delivery service. It was created with small restaurants trying to survive the COVID-19 pandemic in mind.
Tech used: MongoDB Atlas Express.JS React Node.JS Bulma CSS Cloudinary Positionstack Stripe PaymentsDemo Source Code
This was my Fall 2020 COMP562 (Intro to Machine Learning) Project at UNC-Chapel Hill. We used a
pre-trained Multi-Task Cascaded Convolutional Neural Network (facenet-pytorch MTCNN) and trained a
PCA-SVM pipeline to detect faces in images and then classify them to see whether people in those images
are wearing masks.
Dataset: The Face Mask Detection dataset on Kaggle was used - https://www.kaggle.com/andrewmvd/face-mask-detection
Tech used: Python Matplotlib OpenCV Pytorch (MTCNN) SciKit Learn (PCA, SVM) Jupyter Notebook
MediScan is an Android app that allows the user to take a picture of any on-paper medical record. The user will then highlight areas of typed/handwritten text information to then digitize them into easily-readable spreadsheets by using Optical Character Recognition (OCR) via Google Cloud Vision.
Tech used: Google Cloud Vision Android Studio JavaDevPost Source Code
Newstopia is a community-based mobile app that allows its users to critically rate articles and have meaningful, civil discussions regarding media trust and article legitimacy. I came up with the idea and then built a prototype along with my friends at HackNC 2019 - it was an iOS prototype app named "TrustMe". The DevPost link to TrustMe is provided below. Much more is coming to the app, and it is being developed privately.
Tech used: Google Cloud Firestore Algolia Xcode SwiftDevPost (Prototype named TrustMe) Source (still private)
A dashboard that provides insights into Chapel Hill’s crime data, via visualizations made using Tableau.
Tech used: Tableau Wix website builder🧐 Click for pics Project Dashboard
I came up with the idea, designed the project's architecture, was responsible for integrating the software components, and led our team of 5 to a top-10 finish at an all-India 36-hour hackathon with 170+ competing teams. The project consists of a RaspberryPi system connected to a camera, triggers recording upon motion detection, performs face recognition, and sends SMS intruder alerts via the Nexmo API that include a link to the live camera stream which we had set up using FFMPEG, an open-source live streaming solution. Our team received a Rupees 100k (about 1,350 USD) grant offer from VIT University's Technology Business Incubator for the same.
Tech used: Python OpenCV Nexmo SMS API RaspberryPi
A multi-lingual QnA platform for Indian farmers made using PHP, MySQL, Bootstrap and Google Translate API. Features a sentiment Analysis Engine (uses a Naive Bayes classifier) written in Python, and a dashboard made using Chart.js that provides stats about farmers' mental health.
Tech used: Python Naive Bayes Chart.js PHP