Hi, I'm Steven 👋

I-Min (Steven) Chen is a data scientist and AI application developer pursuing an M.S. in Data Science at Vanderbilt University. I build human-centered tools across healthcare, machine learning, and applied research.

🔗 LinkedIn · 📫 [email protected]

Featured project

🩺 Project CRC-ORACLE 腸命天機

An interactive research prototype for exploring outcomes among historically similar colorectal cancer patients. A clinician can describe a patient in natural language, review 23 structured clinical features, locate the patient in a Factor Analysis of Mixed Data (FAMD) space, and compare the selected cohort using Kaplan–Meier survival curves and postoperative outcomes.

• Data: 12,196 de-identified synthetic historical patient records
• My role: Front-end development, clinical UX, natural-language parsing, model integration, validation, and deployment
• Design focus: Transparent defaults, traceable variable mapping, bilingual UI, and clear clinical-use limitations
• Team: StanCode SC201 final project with colorectal surgery and data science collaborators

Open the live demo · JavaScript FAMD Survival Analysis Clinical UX

This prototype presents historical cohort statistics for research demonstration. It is not an individualized diagnosis, prognosis, or treatment recommendation.

Other projects

📱 iOS apps

• ✈️ Packmon — Your travel packing partner. SwiftUI iOS app that builds personalized packing lists and flags customs/airline-security risks before you fly. Team project (TravelGenius), showcased at Taiwan Student Creator Community. Live site Swift SwiftUI

🤖 AI × data

• 🧠 Cz Landmark Estimation for fNIRS Co-registration (Vanderbilt BrainHack 2026) — Developed two complementary methods for recovering the Cz scalp landmark when an fNIRS cap obscures the inion: a C7-referenced physical protocol for new participants and a residual-correction framework for historical data. Kept the current n = 4 analysis descriptive and outlined a path toward PCA, clustering, random effects, and mesh-based shape modeling. fNIRS Neuroimaging Statistical Modeling Experimental Design
• 👶 Parent Navigator 育兒導航全攻略 — AI parenting assistant for new parents in Taiwan. Built as Capstone Project Manager, ITRI AI & Big Data Program: async webhook architecture, RAG pipeline with LangChain, safety guardrails, Supabase backend. Full project · Live demo Flask Gemini RAG Supabase
• 🩺 Nurse Resiliency (Vanderbilt) — Burnout-risk modeling for 6,000+ medical providers using sentence-transformers, clustering, and XGBoost. (methods showcase — data not shareable)
• 📊 Probability & Statistical Inference Portfolio — 8 deliverables from Vanderbilt DS 5620: roulette simulation, Monte Carlo error, order statistics, MLE & method of moments. R Statistics

🌐 Web

• 📮 Post Fare Calculation — Postal fare calculator with a postcard-style UI. JavaScript GitHub Pages

💼 Experience

• Wealth By Health Foundation — Project Manager / Outreach Coordinator (2022–2025): automated SQL/Python data pipelines for 100+ clinic events serving 2,000+ underserved patients annually; forecasting models raised patient throughput 25%
• Fubon Financial Holding — Summer Data Analyst: quantitative modeling (regression, time series, Optuna tuning) for automated trading
• Western Michigan University — Finance Research Assistant: bank evaluation with R and time-series forecasting
• Also: TA for Stanford's Code in Place course · Representative, UN ECOSOC Global Partnership Forum

🛠 Skills

• Languages: Python · R · SQL · JavaScript · Swift
• ML / AI: NLP · transformers · RAG · LLMs · LangChain · XGBoost · scikit-learn · TensorFlow
• Data: Supabase (PostgreSQL) · MySQL · Power BI
• Research: fNIRS · Neuroimaging · Experimental Design · Scientific Writing · Git

🎓 Education

• M.S. Data Science — Vanderbilt University (2027)
• B.B.A. Finance — Western Michigan University · B.A. Economics — National Chung Cheng University (dual degree)