Welcome to the Universe

I'm a passionate, curiosity-driven problem solver exploring the infinite possibilities of our grand universe, creating experiences that reach for the stars and galaxies.

Explore My Work

About Me

Welcome to my corner of the digital universe! I'm a mathematician and physicist by training, a builder by passion, and a dreamer by philosophy. I thrive at the intersection of theory and application, where abstract ideas become real projects that push the boundaries of what's possible.


At the heart of my journey is a simple principle: everything is possible. Whether I'm exploring new frontiers in AI, physics, or business, I see each challenge as an invitation to expand what humanity can do.


My long-term dream is to build personal spaceships and make space exploration accessible for everyone. Until then, I'm building the skills, knowledge, and ventures that bring that future closer, one step at a time.


When I'm not coding or designing experiments, I'm stargazing, learning new languages, or building ventures that test how ideas scale in the real world. Each pursuit, whether scientific or creative, connects back to my lifelong goal: making the improbable possible and guiding curious minds toward their own frontiers of possibility.

Portfolio

Here are some projects I've worked on, each one a small universe of its own:

Stablecoin De-Peg Prediction

Data analytics project forecasting stablecoin de-peg events using machine learning. Built SQL + Python pipeline achieving 95-98% ROC-AUC accuracy.

Python SQL DuckDB Machine Learning
View Project →

YouTube Shorts Extension

Chrome extension converting YouTube Shorts to normal video UI. Featured on Chrome Web Store with 500+ users and 5.0/5.0 rating.

JavaScript Web API DOM Manipulation
View Project →

Astrophysics Research

Processed gravitationally lensed dusty star-forming galaxy (DSFG) data, reconstructed galaxy images, and calculated redshift using Python and specialized astrophysical software.

Python Astropy SAOImageDS9 Astrophysics
View Project →

High Power Rocket Launch

Designed and built H-class motor rocket using CAD, 3D printing, and simulation. Successfully launched to 400m with NAR Level 1 certification.

Fusion 360 OpenRocket 3D Printing Engineering
View Project →

Kyrics - K-Pop Themed Korean Learning Platform

Web service for learning Korean through K-pop lyrics. Led 11-person team to launch MVP, acquired 500+ beta subscriptions.

Product Management Figma Notion Web Development
View Project →

Contact

Ready to embark on a project together? Let's connect and create something extraordinary!


Email GitHub LinkedIn

I'm always open to discussing new opportunities, collaborating on interesting projects, or just talking about the latest developments in technology and space exploration.

Stablecoin De-Peg Prediction

Data Analytics Dashboard

Predicting Stablecoin De-Peg Events

Google Data Analytics Capstone project forecasting stablecoin de-peg events for fictional fintech Oruda's Risk team using machine learning.

Python SQL DuckDB Machine Learning Logistic Regression
GitHub Repository

Overview

This capstone project addressed the critical business need of forecasting stablecoin de-peg events for risk management. I built a comprehensive data pipeline and machine learning model to predict next-day de-peg risks with high accuracy.

Key Achievements

  • Processed 700+ daily observations of stablecoin, Bitcoin, and U.S. dollar index data
  • Engineered features including volatility, returns, and deviations
  • Labeled 40+ historical de-peg events for model training
  • Achieved ROC-AUC of 0.95-0.98 with logistic regression models
  • Proposed >2% deviation alert system for stakeholders

Technical Implementation

Built a robust SQL + Python pipeline using DuckDB for efficient data processing. The machine learning approach focused on logistic regression models optimized for financial time series data, with careful attention to feature engineering and model validation.

YouTube Shorts Extension

Chrome Extension Interface

YouTube Shorts → Normal Video Extension

Chrome extension that converts YouTube Shorts into standard YouTube video interface, enhancing user experience and control.

JavaScript Web API DOM Manipulation
Chrome Web Store GitHub

Overview

This Chrome extension addresses user frustration with YouTube Shorts by automatically converting them to the familiar standard YouTube interface, providing better controls and viewing experience.

Key Achievements

  • Peaked at 4,000 users with 500+ currently active
  • Achieved Featured badge on Chrome Web Store
  • Maintained 5.0/5.0 user rating
  • Passed Google Web Store review process
  • Continuously iterated based on user feedback

Technical Implementation

Built using vanilla JavaScript with Web APIs, focusing on efficient DOM manipulation and seamless user experience. The extension monitors page changes and automatically applies the conversion while maintaining YouTube's functionality.

Astrophysics Research

Astrophysics Data Visualization

Gravitationally Lensed Dusty Star-Forming Galaxy Analysis

Analyzed CO(J=4-3) line detection data from the Northern Extended Millimeter Array (NOEMA) for a gravitationally lensed dusty star-forming galaxy (DSFG) at z = 3.87.

Python Astropy Spectral-Cube SAOImageDS9 Jupyter Notebook lensmodel

Overview

This research focused on processing and analyzing complex 3D data cubes from NOEMA observations of the DSFG ACT-S J0116-0004. The goal was to understand the spatial distribution and intensity of CO emission, reconstruct the pre-lensing galaxy image, and calculate its redshift.

Key Achievements

  • Processed 3D data cube with sky coordinates and velocity channels using SAOImageDS9 and Python.
  • Wrote Python scripts for efficient data management, visualization, and debugging in Jupyter Notebook.
  • Generated integrated zeroth moment maps with contour maps for initial data parsing.
  • Calculated redshift (z) based on CO(J=4-3) rest frequency and observed frequency.
  • Employed `lensmodel` software to "delens" DSFG data, reconstructing original galaxy image and quantifying magnification factor (μ).
  • Developed comprehensive visual representation of research findings, creating a four-panel image showcasing observed moment map, model moment map, residual noise map, and reconstructed map.
  • Collaborated with peers and provided guidance on utilizing Linux environment through Secure Shell Protocol (SSH).
  • Attended productive daily and weekly meetings, fostering open dialogue and discussing future directions.
  • Extensively reviewed pertinent literature on DSFGs, their analysis techniques, and underlying theoretical foundations.

Technical Implementation

The project involved extensive use of Python with libraries like Astropy and Spectral-Cube for data manipulation and analysis. SAOImageDS9 was crucial for initial data exploration and extraction. The `lensmodel` software was employed for gravitational lensing reconstruction, a key step in understanding the true properties of the distant galaxy. Jupyter Notebooks were used for an organized and reproducible workflow, and SSH for collaborative work in a Linux environment.

High Power Rocket Launch

Rocket Launch Photo

MDRA Red Glare 2023 Rocket Launch

Designed, built, and successfully launched an H-class motor rocket at the MDRA Red Glare event in Maryland, earning NAR Level 1 certification.

Fusion 360 OpenRocket 3D Printing Laser Cutting Engineering

Overview

This engineering project involved the complete design, simulation, construction, and successful launch of a high-power rocket. The project demonstrated skills in CAD design, simulation, manufacturing, and real-world testing.

Key Achievements

  • Successfully launched rocket to 400m altitude
  • Achieved full parachute deployment and safe landing
  • Completed damage-free recovery
  • Earned Level 1 High Power Rocketry Certification from NAR
  • Modified parachute deployment system pre-launch

Technical Process

Used OpenRocket Simulator for flight dynamics modeling and Fusion 360 for detailed CAD design. Manufacturing involved 3D printing for custom components and laser cutting for precision parts. The project required careful attention to safety protocols and NAR certification requirements.

Kyrics - Korean Learning Platform

Kyrics Platform Interface

Web Service for Learning Korean with K-pop Lyrics

Product management role leading the development of an innovative language learning platform that uses K-pop lyrics to teach Korean.

Product Management Figma Notion Web Development Team Leadership
Official Notion Page

Overview

Kyrics was an innovative EdTech project that made Korean language learning engaging by using popular K-pop lyrics as learning material. As Product Manager, I led the entire product development lifecycle from concept to launch.

Key Achievements

  • Directed 11-person cross-functional team
  • Launched MVP within five weeks
  • Acquired 500+ beta waitlist subscriptions
  • Conducted comprehensive customer research
  • Optimized landing page for conversion

Leadership & Process

Coordinated workflow using Figma for wireframing and Notion for project management and documentation. Ensured clear communication across design and development teams while maintaining focus on user needs and business objectives. The project demonstrated skills in product strategy, team leadership, and agile development methodologies.