During December 2024, Bitsnaps integrated the llama-3.2-90b-vision-preview AI model provider into the get-convex/chef repository, expanding its supported AI capabilities. They focused on updating TypeScript-based configuration and provider management, ensuring the new model could be easily listed and utilized within existing workflows. This approach emphasized maintainability, allowing future model integrations with minimal code changes. By leveraging full stack development skills and a configuration-driven strategy, Bitsnaps enabled higher-performance AI experimentation while minimizing risk to current systems. The work demonstrated thoughtful engineering depth, addressing both immediate business needs and long-term extensibility for AI model support within the chef platform.