feat: TensorRT-LLM Inflight batching

[tikikun]

Relevant docs can be found here.

https://nvidia.github.io/TensorRT-LLM/batch_manager.html#get-and-send-callbacks

Inflight batching is the most beneficial feature in CUDA system for LLM inferencing right now it can enable very high throughput.

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 15 days."