feat: Variet Engine v1.0 + 5-model tuning complete

Phase 01 (LLM Tuning):
- Gemma4 26B: 74.65 t/s (fast)
- Qwen 35B: 61.62 t/s (balanced)
- Gemma4 31B: 16.0 t/s (deep-coder)
- Qwen 27B: 16.7 t/s (deep-logic)
- Qwen 122B: 8.95 t/s (ultra, GPU 1 only)

Phase 02 (API Engine):
- FastAPI reverse proxy on port 8000
- /engine/switch hot-swap with 503 protection
- config/engine_models.json as single source of truth
- Replaced 4 individual .bat files with unified engine

File cleanup:
- scripts/ 85 files -> 9 + _archive/
- Root .bat files -> _archive/

scripts/_archive/results/122b_ncpumoe_results.json

@@ -0,0 +1,37 @@
+[
+  {
+    "name": "n-cpu-moe=64 (all CPU)",
+    "n_cpu_moe": 64,
+    "speed_tps": 4.87,
+    "vram_gpu0": 4257,
+    "vram_gpu1": 3793,
+    "vram_total": 8050,
+    "status": "OK"
+  },
+  {
+    "name": "n-cpu-moe=56 (8 layers GPU expert)",
+    "n_cpu_moe": 56,
+    "speed_tps": 4.78,
+    "vram_gpu0": 4233,
+    "vram_gpu1": 3793,
+    "vram_total": 8026,
+    "status": "OK"
+  },
+  {
+    "name": "n-cpu-moe=48 (16 layers GPU expert)",
+    "n_cpu_moe": 48,
+    "speed_tps": 4.82,
+    "vram_gpu0": 4233,
+    "vram_gpu1": 3793,
+    "vram_total": 8026,
+    "status": "OK"
+  },
+  {
+    "name": "n-cpu-moe=40 (24 layers GPU expert)",
+    "status": "BOOT_FAIL"
+  },
+  {
+    "name": "n-cpu-moe=32 (32 layers GPU expert)",
+    "status": "BOOT_FAIL"
+  }
+]