AI infrastructure & inference platform engineer. I build the quiet machinery behind intelligent systems: Kubernetes-native LLM serving, GPU platforms and GitOps. 12+ years running Kubernetes across cloud, hybrid and edge.

• LLM serving: vLLM, SGLang, TensorRT-LLM, Triton, KServe, NVIDIA Dynamo and NIM, with normalized latency and KV-cache metrics
• GPU sharing & scheduling: Kubernetes DRA, MIG, MPS, time-slicing, Kueue, KAI Scheduler
• GPU observability & FinOps: DCGM and AMD GPU Operator metrics, idle/unhealthy GPU detection, per-namespace chargeback
• Inference benchmarking: GuideLLM load tests compared across GPU sharing modes
• Platform engineering: Flux CD across 80+ clusters, Rook-Ceph, multi-tenant isolation
• Post-quantum security & edge networking: WireGuard, ML-KEM/ML-DSA, 900+ device fleet

• QubeSec: Kubernetes operator for post-quantum key exchange, signatures and X.509 certificate lifecycle
• Kubernetes LLM Inference Platform: reference architecture (LiteLLM, Envoy AI Gateway, Gateway API Inference Extension, KServe, llm-d, vLLM with DRA, KEDA). A design and write-up, not a claim of running the full stack in production.
• Kubernetes GitOps Platform: public Flux CD bootstrap and reconciliation patterns
• Former Technical Steering Committee member, Magma open-source 5G core

Stack: Kubernetes · Go · TypeScript · vLLM · SGLang · KServe · NVIDIA Dynamo · GPU Operator · Prometheus · Flux CD · Terraform Certified: CKA · CKS · AWS Solutions Architect Associate