George Wang

Data Scientist (Drug Discovery Intern) @Humanwell Pharmaceutical Us, Inc

St Louis, MO, US

Developed a predictive model for Mean Squared Displacement (MSD) using SMILES molecular structures, converting SMILES into Extended Reduced Graph (ERG) fingerprints with RDKit. Implemented an ensemble learning system (SVM, Random Forest, XGBoost, MLP) achieving 60% accuracy in 5-fold cross-validation with limited data (800 samples) and deployed as a FastAPI-based inference API on the company’s internal server, enabling real-time predictions via REST endpoints. Designed and implemented a web-based data visualization dashboard using Python, Streamlit, Pandas, and NumPy, hosted on AWS EC2, automating experimental result recording, statistical analysis, and generating interactive real-time visualizations for researchers. Visualized SAR datasets using dimensionality reduction and clustering techniques with t-SNE and Linear Discriminat Analysis.