feat: add Jetson power mode evidence

[hyeokjun32]

Summary

• Added measured Jetson Orin Nano TensorRT FP16 15W Runtime evidence.
• Added a 15W vs 25W power-mode comparison summary for Local Studio demo evidence.
• Exposed the power-mode summary through /studio/api/demo-evidence and rendered key deltas in Studio.

Validation

• poetry run python3 -m pytest tests/test_studio_routes.py -q
• poetry run python3 -m pytest -q
• Jetson 15W Runtime schema check passed with tests/test_lab_result_schema.py

Impact

• Strengthens InferEdge Jetson Evidence Track with measured power-mode behavior.
• Keeps the evidence scoped to deployment validation rather than production inference serving.

Notes

• This is a short Runtime smoke, not a sustained thermal benchmark.
• AIGuard remains optional; Lab owns deployment decision context.