Zhuoyu Wang

• Collected and processed large-scale text data for analysis, focusing on electric vehicles in the US and worldwide.
• Applied natural language processing (NLP) techniques to extract insights from text data, supporting research on global electric vehicle trends.
• Conducted detailed text analysis to identify key patterns and developments in electric vehicle policies and innovations.
• Collaborated with interdisciplinary teams to analyze text data and generate actionable insights for urban planning and sustainability efforts.

• Developed and deployed machine learning models for real-time eating behavior detection, integrating models into a web-based application.
• Enhanced backend performance and reduced latency for seamless real-time data processing using Python, TensorFlow, and JavaScript.
• Collaborated cross-functionally to optimize model accuracy, implementing data preprocessing and analysis pipelines for improved reliability.
• Contributed to the design and optimization of scalable machine learning solutions, improving app performance and user experience.

• Validated a state-of-the-art transfer learning algorithm and applied it to image classification using logistic regression.
• Improved classification accuracy, earning an A grade for the project.
• Used Python and R for model development, applying transfer learning to enhance model performance.

• Spearheaded the reconstruction of the R package dabestr and developed new features for the DABEST-python package.
• Implemented estimation statistics for effect sizes and confidence intervals in experimental testing, replacing traditional p-value approaches.
• Collaborated with cross-functional teams using GitHub and Basecamp for package development and communication. https://www.claridgechang.net/people.html

• Utilized ResNet and EfficientNet V4 models for pneumonia detection in X-ray images, achieving 99% accuracy.
• Integrated transfer learning with retrained weights from large source datasets to enhance model performance.
• Collaborated virtually using Zoom, Microsoft Teams, and GitHub to develop a comprehensive modeling approach.

• Leveraged deep convolutional neural networks (VGG16, ResNet) to create a backend system for handwritten number detection and recognition.
• Contributed to the development of models and APIs for efficient text image and handwriting recognition

• Employed decision support vector machines, decision trees, and XGBoost models to detect fraudulent activities with 98% accuracy.
• Collaborated within a team of six to research, develop, and validate machine learning models for fraud detection.

• Employed unsupervised clustering methods (k-means, apriori) and decision tree models to analyze customer and geospatial data.
• Predicted customer demand for Grab’s international customer datasets and detected faulty chips using computer vision models.

• Learned to use Auto-CAD, DIA-Lux, Excel for testing light fitting designs, solved a school lighting problem for the company and explored some data analysis with DIA-Lux software for luminosity settings

• Developed an AI-powered chatbot using natural language processing (NLP) techniques to automate resume analysis and candidate screening.
• Implemented algorithms to extract key skills and experiences from resumes, improving candidate selection efficiency.
• Utilized Python and NLP libraries to design a robust system for parsing and analyzing resume data, providing automated insights to streamline hiring processes.

https://github.com/Oliverzhang0314/llm-app-dev-dsa4213