Developed and integrated an ONNX to TOSA conversion feature within the Xilinx/onnx-mlir repository, enabling seamless model conversion and optimization for deployment on target accelerators. The work focused on end-to-end automation, reducing manual intervention and accelerating model readiness for production environments. Leveraging Python scripting, YAML configuration, and Docker-based CI/CD pipelines, the developer ensured robust integration into the main branch, stabilizing the baseline for downstream pipelines and customer deployments. Emphasis was placed on integration stability and validation, with no major bugs reported during the period. This contribution established a repeatable, efficient path for ONNX model optimization and deployment.