Add LLMKube to Inference Platform section

[Defilan]

LLMKube is a Kubernetes operator for llama.cpp-native LLM inference.

• GitHub: https://github.com/defilantech/llmkube
• Apache 2.0 license
• CRD-based model and inference service management
• NVIDIA CUDA and Apple Silicon Metal GPU support
• Multi-GPU layer sharding, pre-flight memory validation
• Helm chart, Prometheus metrics, OpenAI-compatible API

I'm the creator and maintainer. Happy to provide any additional info needed.