MDVax was a website and Twitter alert tool that helped
individuals in Maryland schedule appointments for the
crucial COVID-19 vaccination. Because of the high demand and
low supply, many appointments were initially distributed on
a first-come, first-serve basis, so naturally, these
appointments filled up promptly. The solution I engineered
was a web scrapping system that continuously collected data
on appointment availability and aggregated that data in one
place. When appointments became available, my tool
automatically updated the website and sent out a Tweet
As a result, this project garnered significant press
attention from many local, state, and even national
publications because of its usefulness and the public good.
Linked below are several of the interviews and articles that
have been made about this tool. Notably, the website has
over 250,000 unique users and 3 million page views.
Similarly, the Twitter feed has grown to over 17,000
followers, 10 million impressions, and 1 million profile
visits. The growth and support I received from this project
were humbling and inspiring because of all the gratitude
people have sent after successfully using my tool. I proudly
accepted donations on behalf of a few essential charities,
and we’ve raised over $20,000 for various causes.
My startup, crepkitchen, was the solution to the constant
issue of customers missing out on limited-edition clothing
and sneakers. We created an online subscription service that
instantly notifies users of product restocks and other
essential information for limited-edition sneakers and
clothing releases before they sell out. The notification
then allows the customers to succeed during limited-edition
releases to either flip the item for a profit or collect
The idea started to help some friends and myself get our
shoes faster and cheaper, which became the basis of the
startup founded in my senior year of high school. From
pitching as a finalist in front of hundreds to the entire
University of Maryland’s Pitch Dingman Competition to
meeting with angel investors, crepkitchen was my first
glimpse into the startup world, which I am eager to explore
more of in the future..
• $15,000 MRR (monthly recurring revenue)
• 300 MAU (monthly active users)
This personal project started while I was learning more
about technical analysis of stock charts. I used some of the
strategies I learned and programmed them by creating simple
indicators using quantitative finance formulas. Soon after,
I found I had a data set of several indicating statistics on
all popular stocks, so I decided to implement them into a
machine learning algorithm.
You can read more about this project via a published
explanatory blog post I wrote below or from this project's
GitHub page. This was my first blog post and garnered a lot
of positive feedback.
Disclaimer: This is not financial advice and should be
used for educational purposes only.
Please reference the blog post and github page for further
This project, called Oravise, was a web platform
brainstormed, designed, and engineered at the University of
Maryland’s hallmark hackathon, Bitcamp, in April 2019.
Oravise was a peer to peer scholarly website connecting
individuals curious about unsolved world problems across
academic fields, including astronomy, computer science,
mathematics, medicine, and other essential subjects with
researchers in those respective fields. The website allows
users to select an academic researcher from a list given the
users chosen subject, subcategory, and unsolved problem and
connect with them. I worked with a small team to build this
project and learned everything from Flask to learning new
APIs to making my own API and much more.
Another hackathon project called Election in Tweets and was
made at Vandy Hacks 2018. My teammate and I have always been
big fans of Nate Silver’s website FiveThirtyEight where he
predicts winners of political elections. We decided one data
set not used that often was social media analysis.
To analyze every tweet to predict this election, I had to
build my own Twitter API and save every tweet. Then ran this
data set through sentiment analysis to gauge the popularity
of specific candidates. The data was then compiled, and by
each state, we graphically represented our predictions based
on the model.
Our model performed better than expected and even
outperformed FiveThirtyEight by accurately picking over 95%
of the winners based on just social media interactions.