Worked on expanding the evaluation capabilities of the EvolvingLMMs-Lab/lmms-eval repository by integrating support for AWS Bedrock and local OpenAI-compatible LLM providers. Leveraged Python and API integration skills to update the ProviderFactory, enabling seamless registration and usage of new providers within evaluation workflows. Introduced .env-based configuration loading using python-dotenv, which streamlined deployment and authentication processes, particularly for handling bearer tokens with Bedrock. This feature broadened the framework’s ability to evaluate both cloud-hosted and self-hosted models, enhancing scalability and flexibility for diverse model ecosystems while laying a foundation for more reproducible and configurable evaluation pipelines in backend development.