StarVLA: A Lego-like Codebase for Vision-Language-Action Model Developing | Deep Tech Hub | Startup Networx