Ridge Chang Résumé

I grew up in Savannah, Georgia, and did both my undergrad and my MSBA at Emory University. I learn best by doing. When something’s new to me, I start messing with it, pull it apart and put it back together, and that’s how I approach data too.

Jan – May 2026

Data Science Practicum

Cox Automotive · Goizueta MSBA

Built a B2B customer churn model on a four-person team using an Explainable Boosting Machine with expanding-window temporal cross-validation, optimized for PR-AUC and top-decile lift. Findings informed account-management prioritization for at-risk customers.

Aug 2024 – May 2025

LLM Research Assistant

Emory University · AI/NLP Lab

Built a RAG system over FDA regulatory meeting documents, reaching ~80% retrieval accuracy through iterative chunking and embedding refinement. Collected and processed FDA advisory committee transcripts for an ongoing study on predicting medical product efficacy, engineering NLP features that quantify how the committees deliberate.

Aug 2024 – Present

Hackathon Operations Director

Emory Hacks · MLH-affiliated

Run end-to-end operations for Emory’s flagship hackathon: 208 participants from 35 universities and 40 projects in 2025, a $10K+ budget, and sponsorship tiers from $500 to $3,000.

May 2022 – May 2024

IT Consultant & Assistant Manager

Computer Networking Resources · Savannah, GA

Managed $100K+ in monthly procurement and supervised a team of nine, improving order processing times by 50% through restructured vendor communication. Assisted 100+ clients in manufacturing, legal, finance, and healthcare with computer and network issues.

  • FDA efficacy prediction

    2025

    Predicting medical product efficacy from the quality of FDA advisory committee deliberations, using NLP features built from meeting transcripts.

  • FDA regulatory RAG

    2025

    Retrieval-augmented generation over FDA meeting documents. About 80% retrieval accuracy after iterating on chunking and embeddings.

  • LLM detection breaking

    2026

    Experimented against state-of-the-art AI-text detectors and found a consistent method that lowered detection rates by upwards of 21%.

  • LLM jailbreaking experiments

    2025

    Tested jailbreak methods against frontier models, including GPT 4.1, Grok, Sonnet 4, and several image models.

2025 – 2026

M.S. in Business Analytics

Emory University · Goizueta Business School

Machine Learning, Deep Learning, Causal Inference, Prescriptive Analytics, Statistics, Computational Linguistics, Big Data & Cloud Technologies, Economics of Digital Platforms

2021 – 2025

B.B.A., Information Systems & Operations Management

Emory University · Minor in Computer Informatics

Languages & querying
Python (pandas, scikit-learn, PyTorch, PySpark), R, SQL, Java
ML & stats
XGBoost, LightGBM, CatBoost, InterpretML (EBM), MLlib; causal inference (PSM, IPTW, IV, A/B testing); imbalanced classification (SMOTE, PR-AUC, threshold tuning)
LLMs & NLP
LangChain, LangGraph, Hugging Face, RAG pipelines, fine-tuning
Data & cloud
Snowflake, BigQuery, MySQL/AWS RDS, Spark/Hadoop, AWS, GCP, Azure, Tableau, Power BI