[Demo] Create the canonical three-step Model Connect demo

[yifeif-nv]

Summary

Create one polished demonstration that communicates the core TensorRT Model Connect experience:

1. Build a supported model.
2. Run the generated bundle.
3. Produce an observable result.

Tasks

• Select and pin a representative model and checkpoint with reliable setup, build, and runtime behavior.
• Define the supported hardware, operating system, TensorRT Model Connect revision, and dependency versions.
• Keep the primary path to the minimum required commands.
• Show the expected bundle artifact and observable inference output.
• Demonstrate that the deployed native runtime does not require Python or PyTorch.
• Include approximate setup, build, and inference times while identifying them as hardware-dependent.
• Provide copyable commands and troubleshooting guidance for common failures.
• Prepare both a live-demo path and prerecorded fallback assets.
• Validate the complete workflow from a clean environment.
• Add an automated smoke check for the documented command and output contract where practical.
• Link the demo from the repository and documentation.
• Package the material so it can be reused for talks, livestreams, and technical demonstrations.

Acceptance criteria

A new user can reproduce the complete build, bundle, native-run, and observable-output workflow from the published instructions in a clean supported environment.

The same validated material can be delivered live or prerecorded without changing the documented technical path.

[T4t4KAU]

Hi, I'd like to help with #1107. Would it be useful if I take a scoped first slice covering the canonical Qwen E2E path: pinning the validated environment, verifying build → bundle → native-runtime execution from a clean environment, and adding a smoke check for the documented command/output contract? I’d keep the PR focused and leave the broader demo assets/live-demo packaging separate.

[yifeif-nv]

Hi, I'd like to help with #1107. Would it be useful if I take a scoped first slice covering the canonical Qwen E2E path: pinning the validated environment, verifying build → bundle → native-runtime execution from a clean environment, and adding a smoke check for the documented command/output contract? I’d keep the PR focused and leave the broader demo assets/live-demo packaging separate.

Hey @T4t4KAU thanks for the offering and sorry for the delayed reply! Unfortunately, I haven't really got a chance to break down the task and we are looking for a demo that is a little more then just the build -> bundle -> runtime. i.e. building a simple application on top of model connect. Let me know if you have some good ideas, and we can create a sub-issue and working on such example that is attractive to people