Computer Scientist | BU '25 BA/MS (Dual Degree)
Simon Kye
Portfolio
Hi! My name is Simon Kye. I'm a computer scientist studying in Boston Unviersity completing the MS portion of my BA/MS dual degree. (The BA portion has been finished and I'm able to get my bachelor's at any point)
I'm a Korean who grew up in Hong Kong that is now studying in the United States. This has allowed me to experience multiple varying cultures and helps me adapt to many foreign scenarios.
I love questioning the status quo and constantly asking if something is good enough. For example, I knew my school's course search could be optimized in many ways and decided to make a optimized course search for fellow BU students.
What motivates me is simply being able to see the impact and the scope of problems I can solve through programming, something I used to do for fun as a kid. I've worked on allowing easier access to criminal justice data to help under-represented groups have a fairer trial. I've helped missing people be identified easier and faster through the use of machine learning. I'm excited to see what else I can contribute to with the skills that I have.
Created a Jupyter Notebook for a client that uses the encoder-decoder method to apply the effects of homelessness to photos of missing people to assist in identification of said missing people. Uniqueness of the problem given required reading through many scholarly research papers and testing the models used in said papers to find the appropriate model that would meet the requirement of the task at hand.
Created an Android app using Kotlin & Jetpack Compose which was designed in Figma with Google's M3 design template. The app allows users to post items that they've thrown away at the curb such as a couch so that other users can pick them up. This app supports filtering, searching, ratings, and ensures users can only claim items if they're close enough to the posted items.
Using OCR, this app allows users to take a picture of their receipt. The picture is then processed and will return a intuitve interface for users to split their bill. The user can then share the result of the split bill.
Inspired by Boko33's portfolio website, I decided to make my own room in Hong Kong as I'll be moving from a room I lived in for 13 years. This was good opportunity to not only learn something new but also create a personal record of where I grew up. All assets were designed from scratch.
A site called Woke Windows provided information about officers and their activites in Boston. This was so that lawyers and advocates could easily access data that could show that certain officers were treating a group of people in a certain way. It has since been shut down, and BU was asked to create a replacement. As part of the core infrastructure team, we were tasked with building the backbone of the site and communicating to all other groups, which were in charge of data collection, about the type of data they could give us to create proper reuseable components. With the given data, we were also tasked with moving it to the database. The site keeps track of all recent boston police activites.
BU's official course search has many unintuitive downsides. It
requires constant opening of new tabs, lacking proper quality of
life filters, and no review/difficulty tracker for the school.
Rather than waiting for the website to get better, I decided to
make a better version myself. I scraped the entire course search
database, and uploaded it to a seperate firebase database to
enable better filters. This database is automatically updated
every 24/7 hours with the use of AWS lambda functions.
BU
Course Search has since replaced and updated their course search
and the site now shows legacy data.
Inspired by GeoGuessr, me and three others created a website. We
practiced making API calls, decoupling the frontend and backend,
and implementing login authentication. The website would display a
artwork grabbed from a mueseum API and the user can choose where
the artpiece originates from. User's accuracy is tracked by using
their google accounts.
Google's OAuth changes has since
broken the site.
Research Assistant
Conducted literature reviews and analyzed relevant studies to support research projects on data regulation, generative AI, trend analysis algorithms, and other topics regarding AI's use cases and role in business analytics. Helped with data collection (usually each "job" taking up over 200GB of data) from Wikipedia & Reddit to assist in papers regarding lifespan of data for AI.
Process Automation Intern
Created automation processes using UiPath to reduce long and repetitive tasks that would require 10+ hours to simple automation processes done in less than a minute. Self-taught UiPath for 50 hours and earned the “UiPath Certified Advanced RPA Developer” certificate. Created automations using computer vision and excel sheets to create summaries of thousands of data points into emails or excel sheets
CS Tutor
Independently taught high schoolers AP CSA/IB CS HL weekly 1-on-1 and basic prgramming to children. Developed a 10-week long curriculum with problem sets, lesson plans, and curated projects based on individual interests aimed at cultivating an interest and passion for computer science, resulting in a student declaring their major in Computer Science.
Teacher's Assistant
Taught high level AMC to 20 students for an hour and twice a week virtually through zoom groups. Created 16 problem sets and solutions using LaTEX that would be easily understood by middle schoolers.