Worked on stabilizing LLM integration and improving Azure OpenAI provider reliability within the Skyvern-AI/skyvern repository. Addressed a configuration issue by updating the model name in the Python-based backend, changing it from AZURE_OPENAI_GPT4O to AZURE_OPENAI in the CLI setup script. This adjustment ensured that the Azure LLM provider was referenced correctly and allowed for proper key-based configuration management. The fix reduced the risk of misconfiguration during deployment and enhanced the reliability of LLM provisioning for Azure OpenAI. The work demonstrated attention to backend development and configuration management, focusing on targeted, high-impact improvements rather than feature expansion.