Routing coding agent requests to the right model on Amazon Bedrock with NVIDIA NeMo Switchyard

NVIDIA NeMo Switchyard, an open source library, lets a coding agent pick which model handles each request, with LiteLLM integration supporting Stage routing (choosing between two candidate models based on recent tool activity) and Random routing for traffic splits and A/B tests. A sample deploys LiteLLM with the Switchyard plugin on Amazon ECS Fargate via AWS CDK, calling models on Amazon Bedrock, and measures cost per solved task against single-model baselines. The sample uses switchyard-server 0.3.0 and the LiteLLM v1.102.0 image, and is for non-production evaluation only.