AI engineer developing smart insights using LLMs to help patients better understand their glucose trends.

• Developed a multi-threaded, API-driven model evaluation pipeline in Python for data fetching, preprocessing, and evaluation, expanding evaluation metrics, and automating processes to reduce time from hours to minutes.• Containerized the evaluation pipeline in Docker and created a RESTful APIs interface using FastAPI, allowing model developers to test model performance via API calls.• Created an automated data visualization engine in Python using both Kibana dashboard on the ElasticSearch and the Seaborn package.• Developed a dynamic frontend interface utilizing React.js and JavaScript, and implemented a fuzzy search feature to enhance data query accuracy and efficiency, improving user experience with comprehensive social media analytics.

- Develop a backend system using Python and Socket.IO to facilitate real-time multi-client connectivity, enhancing user engagement and data exchange.- Ensure robust server performance in multi-client environments, particularly for database connections initiated through Unity applications.

• Developed a cross-platform Group Therapy App using Flutter, enhancing user interaction by enabling profile management and therapy group matching for patients and providers.• Engineered backend APIs using Spring Boot with Java and MongoDB, allowing user login authentication using JWT, profile management, and therapy group information management.

- Developed a remote rehabilitation software interface for stroke patients using Unity, enhancing patient access and recovery prospects.- Integrated a model of the Phantom Omni Touch Robot into Unity using C#, facilitating intricate interaction scenariosin.- Enabled real-time force feedback from the robot, enhancing system responsiveness and user experience.- Leveraged multiple machine learning models for recovery data extraction and analysis, driving personalized and data-driven rehabilitation plans.