Worked on enhancing the google-ai-edge/model-explorer tool by focusing on robustness, usability, and deployment flexibility. Implemented safe import handling to make PyTorch an optional dependency, allowing the explorer to function even when PyTorch is not installed. Improved the command-line interface by supporting multiple model path inputs, including argument separation and glob patterns, which streamlined experimentation and reduced setup friction. Strengthened input validation and startup reliability to facilitate broader adoption without altering downstream usage. Updated documentation to reflect these changes, ensuring clarity for users. The work leveraged Python, backend development, and argument parsing skills to deliver a more accessible and resilient tool.