Detect and identify different species of harmful algae within natural water in real-time with AI and a camera (i.e., ESP32-CAM, smartphone, or webcam).

• Designed and developed a system that uses custom AI (i.e. computer vision) and a camera to quickly detect and identify species of harmful algae in real-time (livestream/video); i.e. object detection using convolutional neural networks (CNNs)
• Integrated various fine-tuned algae detection models (i.e. custom AI) with (1) ESP32-CAM (low-cost development board with camera, WiFi, and Bluetooth modules) attached to a Nikon microscope's eyepiece via a 3D printed lens attachment, (2) iPhone, and (3) webcam as camera options
• Lead and managed team of 3 senior computer science undergrads
• Wrote in-depth, beginner-friendly Google Colab tutorial that details the entire process of training, validating, inferencing, exporting, and deploying the custom AI model; tutorial also goes over dataset annotation, preprocessing, and augmentation

• Created n-qubit quantum program searching for m target(s) in an unsorted database
• Resulted in quadratic speedup over classical/naïve search algorithms

• Simple web scraper to download media from websites
• Supports .pdf generation and vertical image stacking
• Useful for downloading manga, comics, etc.

macOS Sequoia-inspired personal website

Includes (but not limited to) the following features:

• Mail
• iTerm
• Safari
• Notes
• TextEdit
• Launchpad
• Calculator
• Menubar
• Dock
• Desktop
• Music Player
• Trash

• The ideomotor effect of software
• Open-source software project

Human-Computer Interaction project

Developed front-end of web app for BIM models that allows users to:

• Search models for specific elements/properties
• View and inspect models in 2D and 3D
• Query models via virtual assistant (Siri or Alexa)

• The goal of this project is to practice UDP socket programing and understand binary packet structures by developing a simplified DNS lookup client
• You must create your own socket
• You cannot use any existing DNS library

• This project develops a web status monitor (simplified version of uptimerobot.com) to practice web programming and understand the web related protocols: HTTP and TLS/SSL
• Implements HTTP client socket to interact with the web server
• Doesn't use any existing HTTP client library
• Uses an existing SSL library to help implement the HTTPS client (extra credit)

• This program counts the number of prime numbers between 1 to MAX_CONST inclusive using MPI (parallel programming)
• Each process tests its share of the cases and keeps a running total
• Before it finishes, the process prints out its ID number and count of primes that it found
• Master process ends after printing the total number of primes between 1 and MAX_CONST inclusive and the total amount of time taken to find all the primes

• Sorts all words in an input file (or from standard input) and prints the sorted words to an output file (or standard output)
• Command line arguments: bstsort [-c] [-o output_file_name] [input_file_name]
• If -c is present, the program needs to compare the strings in upper case; otherwise, the case stays as read in
• If output_file_name is given with the -o option, the program will output the sorted words to the given output file; otherwise, the output shall be the standard output
• Similarly, if the input_file_name is given, the program will read from the input file; otherwise, the input will be from the standard input

• This program is an extension of myshell.c with pipelines + I/O redirection functionality
• > Redirect standard output from a command to a file; if the file already exists, it will be erased and overwritten without warning
• >> Append standard output from a command to a file if the file exists, else create a new one
• < Redirect standard input from a file to a command
• | Pass the standard output of one command to another for further processing

Program I wrote in high school that converts a given integer to little-endian and/or big-endian

Implementation of multi-party computation (MPC), a quantum-resistant cryptographic algorithm, with custom client-server software to securely compute the product of N massive matrices

1. Split each of N massive matrices into submatrices
2. Homomorphically encrypt the submatrices and disperse evenly among M Internet of Things (IoT) devices
3. M IoT devices calculate the product of N homomorphically encrypted submatrices and send the result back upon completion
4. Received submatrices (i.e. results) are decrypted and concatenated into a single matrix, which is the final result

Some of my solutions to certain CTF puzzles/problems

Shell script to automate the following:

• Create Conda environment(s)
• Remove Conda environment(s)
• Rename Conda environment
• Copy Conda environment
• Export Conda environment
• List Conda environment
• Activate Conda environment
• Deactivate Conda environment

My personal blog

This is a fork of jekyll-theme-chirpy that has been heavily modified to include:

• Graph view plugin
• Tabbed container plugin
• Gallery slideshow plugin
• Link preview
• Image proxy (e.g. for Twitter/X links)
• Jekyll plugin to fix malformed Markdown links
• Auto-generates CNAME and .nojekyll files during build and deployment
• Custom "Important" prompt
• Custom details styling
• And more

iOS app that translates between languages and allows users to store and delete their translations

• Users open the app to a TranslationMe home page with a place to enter a word, phrase or sentence, a button to translate, and another field that is initially empty
• When users tap translate, the word written in the upper field translates in the lower field
• Translation history is stored (in a scroll view in a new screen)
• Translation history can be erased
• Variety of choices for the languages
• Button to conveniently swap languages

iOS app that challenges users with trivia questions spanning a variety of categories, formats, and difficulties

• App launches to an Options screen where user can modify the types of questions presented when the game starts. Users can choose:
  • Number of questions
  • Category of questions
  • Difficulty of questions
  • Type of questions (Multiple Choice or True/False)
• User can tap a button to start trivia game, which presents questions and answers in a Card view
• Selected choices are marked as user taps their choice
• User can submit choices and is presented with a score on trivia game
• Timer that puts pressure on the user; choices are auto submitted after the timer expires

iOS app that allows users to play a simple memory game

• App loads to display a grid of cards initially placed face-down
• Tap cards to toggle their display between the back and the face
  • Tapping a facedown card flips it to reveal the front
  • Tapping a second card that is not identical flips both back down
• When two matching cards are found, they both disappear from view
• User can reset and start a new game
• User can select number of pairs to play with

Simple iOS app that tracks expenses

• User can track their expenses
• User can view home page (with home icon/button at the bottom)
• User can add expenses/etc. by pressing the Plus (+) icon/button at bottom
• User can sign up for an account
• User can log into an account
• User can choose between light and dark mode
• User can view generated graphs based on their data
• User can view and modify profile and/or settings page (also has icon/button at bottom)
• User can separate expenses by category (e.g. food/drink, traveling, entertainment, bills/taxes, etc.)
• User can view overview page with [circle] graphs displaying expenses (also has icon/button at bottom)
• User can track and see how expenses differ from one week/month/etc. to another (e.g. mean expenses on food, gas, etc.)

Archive containing my code for competitive programming problems during high school.

Collection of my config files.

My submission for Teiko's technical exam

File cell-count.csv contains cell count information for various immune cell populations of each patient sample. There are five populations: b_cell, cd8_t_cell, cd4_t_cell, nk_cell, and monocyte.

Each row in the file corresponds to a biological sample. The file also includes sample metadata such as sample_id, indication, treatment, time_from_treatment_start, response, and gender.

Bob Loblaw, a drug developer at Loblaw Bio, is running a clinical trial and needs your help to understand how his drug candidate affects immune cell populations. Your job is to:

• Design a Python program that meets Bob's analytical needs, as outlined in Parts 1-4 below.
• Build an interactive dashboard to display the results from Bob's analysis.

This is the database schema I used:

Collection of all the Tumblr themes I developed while in middle and high school.