Projects

A collection of projects demonstrating technical skills across full-stack development, AI integration, data engineering, and product development. Each project highlights end-to-end implementation from concept to deployment.

Full-Stack & AI

stratLens: AI Strategic Analysis Platform

Personal Project | January 2026 | Full-Stack Development & AI Integration

Project Overview

AI-powered strategic analysis platform that transforms company data into professional strategic analyses with consultant-grade visualizations. Enables strategists and business leaders to generate comprehensive analyses for any publicly traded company in minutes.

Technical Implementation

  • Built full-stack application with Next.js 14 (App Router), TypeScript, and PostgreSQL with Prisma ORM
  • Integrated multiple AI providers (OpenAI, Anthropic Claude, Google Gemini) with BYOK (Bring Your Own Key) architecture
  • Implemented 11 strategic frameworks: SWOT, Porter's Five Forces, PESTEL, BCG Matrix, Value Chain, VRIO, Business Model Canvas, McKinsey 7S, Ansoff Matrix, Three Horizons, Scenario Planning
  • Built real-time data aggregation from Yahoo Finance, News API, and Alpha Vantage for live market intelligence
  • Created consultant-grade visualizations using D3.js and React Flow with PDF/PowerPoint export capability
  • Implemented secure authentication with Clerk and AES-256-GCM encryption for user API keys

Technology Stack

Next.js 14React, TypeScript
PostgreSQLPrisma ORM
Multi-AIOpenAI, Claude, Gemini
D3.jsVisualizations

Skills Demonstrated

Next.js • TypeScript • PostgreSQL • Prisma • AI API Integration • D3.js • React Flow • Clerk Auth • Tailwind CSS • Vercel Deployment

Hedwig: AI-Curated Daily News Briefing

Personal Project | February 2026 | Python & AI Automation

Project Overview

Personal daily news briefing system that fetches, curates, and emails a 20-30 minute digest covering tech, politics, military, business, philosophy, history, and art from trusted sources every morning. Includes source reliability ratings and live market data.

Technical Implementation

  • Built Python-based data pipeline aggregating content from 20+ trusted sources (BBC, Bloomberg, Hacker News, The Economist, etc.)
  • Integrated Claude API for intelligent article summarization and curation
  • Implemented source reliability indicators based on Media Bias/Fact Check and AllSides ratings
  • Added live market data integration (S&P 500, NASDAQ, Gold, Bitcoin) via Yahoo Finance API
  • Built Jinja2-templated email system with mobile-friendly executive-style formatting
  • Automated daily delivery via GitHub Actions with 95%+ test coverage using pytest

Technology Stack

PythonCore language
Claude APIAI curation
GitHub ActionsAutomation
95%+Test coverage

Skills Demonstrated

Python • Claude API • RSS/API Integration • GitHub Actions • pytest • Jinja2 • Data Aggregation • Email Automation

Product & Healthcare

Healthcare ER Length-of-Stay Optimization Study

Los Angeles General Medical Center | January - April 2024 | Data Science & Operations Research

Business Challenge

Hospital emergency department faced operational inefficiencies leading to long patient wait times and bottlenecks. Leadership needed data-driven insights to optimize patient flow and reduce length-of-stay while maintaining care quality.

Approach

  • Preprocessed and cleaned 80,000+ EMR patient records using Python (pandas, NumPy)
  • Built discrete-event simulation model using M/M/2 queuing theory modeling patient flow across 5 triage-to-discharge stages
  • Identified peak congestion periods (9 AM - 1 PM) and critical workflow bottlenecks (decision-to-depart stage)
  • Developed Power BI dashboard tracking throughput, wait times, and provider utilization

Business Impact

80K+records analyzed
12%projected LOS reduction
5process stages modeled
Adoptedby hospital operations

Analysis was adopted by hospital operations team. Recommendations informed staffing decisions and process improvements during peak hours.

Skills Demonstrated

Python (pandas, NumPy) • Queuing Theory • Simulation Modeling • Power BI • Process Analysis • Healthcare Operations • Data Visualization

SwiftER: AI-Powered Patient Management Platform

USC Min Family Challenge for Healthcare Innovation | November 2023 - April 2024 | Product Management

Business Challenge

Healthcare providers struggle with inefficient patient management, manual lab order processes, and poor traffic prediction. Needed AI-assisted platform combining intelligent lab guidance, predictive analytics, and HIPAA-compliant communication.

Approach

  • Led 5-person team through 6-month product development cycle
  • Conducted user research with 20+ healthcare professionals across 12 thematic areas
  • Performed competitive analysis on 10+ EHR systems; built TAM/SAM/SOM financial models
  • Established partnership with Sakura Hospital (Japan) for validation

Business Impact

Semi-finalsMin Family Challenge
20+professionals interviewed
10+competitors analyzed
Japanhospital partnership

Advanced to semi-finals in USC Min Family Challenge for Healthcare Innovation. Validated product-market fit through international hospital partnership.

Skills Demonstrated

Product Management • Team Leadership • User Research • Competitive Analysis • Financial Modeling • Partnership Development • Healthcare Tech • GTM Strategy

Learning Projects

Academic coursework and skill-building projects.

ML Classification, Regression & Clustering

NUS ISS Machine Learning Module | January 2026 | Academic Coursework

Project Overview

Machine learning coursework covering supervised and unsupervised learning paradigms. End-to-end ML workflow from data exploration to model evaluation.

Projects Completed

  • Earthquake Damage Classification: Built model to predict damage grades; performed feature engineering on seismic data
  • Insurance Cost Regression: Developed regression model predicting premiums; achieved strong R² with regularization
  • Customer Segmentation: Applied K-Means clustering for marketing strategy insights

Technology Stack

Pythonpandas, NumPy
scikit-learnML models
matplotlibVisualization
JupyterNotebooks

Daily Diary: Full-Stack Personal Journal

Personal Learning Project | January 2024 | Full-Stack Development

Project Overview

Personal journal application built to learn end-to-end full-stack development. Includes user authentication, full-text search, and responsive design.

Technology Stack

Next.js 14App Router
TypeScriptType safety
PostgreSQLPrisma ORM
NextAuthAuthentication