Worked on expanding performance testing coverage for the TensorRT-LLM repository, focusing on the Llama-3.1-Nemotron-8B-v1 model. Developed and integrated new performance tests across both PyTorch and TensorRT backends, targeting a range of input and output lengths to capture comprehensive latency and throughput metrics. Enhanced the test configuration by including the model path, which supports repeatable and consistent performance runs. Leveraged Python and YAML to implement these changes, utilizing CI/CD practices to ensure reliable integration. This work improved the detection of performance regressions and provided clearer benchmarking data, supporting faster optimization cycles and more robust model deployment readiness.