Bolder-I
- Designed and implemented a file extraction tool for an AI-powered web interface, enabling the processing and analysis of medical client documents (TXT, DOC, DOCX)
- Developed a custom Tools class in Python, integrated into Open WebUI, to support file listing, subfolder navigation, and content extraction, with robust error handling and support for large-scale medical document parsing
- Deployed and managed the Open WebUI application using Docker, configuring containers to mount local directories for seamless file access, ensuring scalability and persistence in a containerized environment
- Utilized Docker commands to install dependencies within the container, stop, remove, and recreate containers with updated mounts, and troubleshoot file permission issues, ensuring reliable operation of the file extraction tool
- Configured system prompts in Open WebUI to enable AI models to interact with the Tools class, allowing users to list files, navigate subfolders, and read medical documents via natural language queries (e.g., “list files in subfolder1”, “read Patient2.txt”)
- Collaborated with the Bolder-I team to test and validate the tool’s functionality, ensuring accurate file parsing and AI-driven insights from medical records, which improved the startup’s ability to process and analyze client data efficiently
- Documented the setup process, including Docker container creation, Tools class integration, and system prompt configuration, to facilitate onboarding and future development for the team
