Hello, I'm Qianjin (Zac) Zhou

Advanced credit risk assessment system analyzing 6M customer transactions to enhance traditional credit scoring methods. Achieved 98.7% classification accuracy across 50+ transaction categories.

• Implemented state-of-the-art NLP techniques with BERT and Llama models
• Engineered 40+ behavioral features achieving ROC-AUC score of 0.87
• Optimized model inference latency while maintaining high accuracy

Built an intelligent recommendation engine analyzing 2.5M Google reviews and 21K business records to predict and suggest personalized business recommendations.

• Developed hybrid recommendation system combining content-based and collaborative filtering
• Achieved 64% improvement in prediction accuracy (MSE: 0.374)
• Implemented sentiment analysis for nuanced review understanding

Developed a scalable cloud-based system for processing and analyzing large-scale e-commerce data (33GB) with optimized computational efficiency.

• Achieved 3000% performance boost using parallel processing
• Implemented Decision Tree model with 88% prediction accuracy
• Reduced AWS operational costs by 40% through optimization