Currently, I'm a college student studying computer science with industry experience working with web development and personal projects on machine learning. My specialties include coding in Python for things from backend APIs to ml experiments. On the frontend, I have experience with both React and Angular to create functional, beautiful apps. Please keep reading to see my skills and projects.
What I do
I like to code things from scratch to get a clean slate, and I pay close attention to creating thoughtful, mobile-first, responsive web apps.
Frameworks and tools:
For this, I endeavor to keep things as simple as possible. I aim to make backends that are secure, efficient, and logical. And I create RESTful APIs that make sense.
Classes to Calendar
This is a web application designed to take an image of a UMN class schedule and extract the class data out of it. It uses Google's Tesseract OCR to parse the text from the image into a calendar event. The user can then add these classes to their Google Calendar on either a new calendar or an already existing one. In the future, this will hopefully be able to work not just for UMN classes, but for any class schedule using NLP. It is a React App with a Flask backend API. The demo website can be found here.
This idea is to create a music journaling app that would allow musicians to easily keep track of practice by setting short- and long-term goals and other features such as inserting recordings inline with text. Additionally, users will be able to share their notes with others and teachers could keep track of their students in a much more timely manner. In the future, too, I hope to allow musicians to analyze their recordings based on certain metrics to track improvement (i.e. intonation, rhythm, etc.). It is an Ionic app using Angular with a Django backend. Find the web app here.
With this project, I endeavored to apply my knowledge of machine learning to some creative end. It is a python implementation of the paper "A Neural Algorithm of Artistic Style" by Gayts et al. The system uses feature representations of images in a neural network to separate and recombine the content of one image with style of another image. Essentially, to create the combined image, a loss function is devised which accounts for the content of an image and the style of another image, and the script finds the image that minimizes this loss function. The above image is a gif of the minima of the loss function at different iterations, showing a french horn in the style of Starry Night by Van Gough. The project uses Keras with Tensorflow backend.
This is a calendar web application built from scratch using a Django backend. The website is a calendar in which members of the community are able to add events to the calendar without directly interacting with the the people in charge. It was inspired in design by Google Calendar. The site is currently not in use as it was merely a project to help me learn web development, but the finished product is hosted on pythonanywhere (without the email functionality since the domain has expired). The calendar has month, week, and day views, the ability to organize events under calendars and show/hide the events in the certain calendar, and a fully-functioning account system using Django.