feat: balance Llama 24K context with concurrent BGE

Adjusted VRAM allocation for concurrent operation:
- Llama: 80% VRAM, 24576 context
- BGE: 8% VRAM
- Total: 88% of 24GB RTX 4090

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

services/vllm/config_llama.yaml

@@ -2,8 +2,8 @@ model: meta-llama/Llama-3.1-8B-Instruct
 host: "0.0.0.0"
 port: 8001
 uvicorn-log-level: "info"
-gpu-memory-utilization: 0.95
-max-model-len: 32768
+gpu-memory-utilization: 0.80
+max-model-len: 24576
 dtype: auto
 enforce-eager: false
 enable-auto-tool-choice: true