Specialized Consulting & Advisory
Enterprise Kubernetes Platforms & Private AI Orchestration.
I collaborate with engineering teams to extend the Kubernetes control plane, run stable clustered databases, and build self hosted LLM serving infrastructures. By designing declarative, clear platforms, I help you secure complete control over your systems and cloud spend.
Pillar 01
Kubernetes Platforms & Custom Operators
Extend the core Kubernetes control plane and build custom system integrations at the container runtime, storage, and networking layers.
Custom Operator & Controller Engineering (Go)
I design and write custom Kubernetes Operators and Custom Resource Definitions (CRDs) in Go using KubeBuilder and Controller Runtime. These controllers automate complex application lifecycles, manage backups and scaling, and synchronize external cloud resources directly through the native Kubernetes API.
Go (Golang)KubeBuilderController RuntimeCRD DesignOperator SDK
Low Level System Interfaces (CRI, CNI, CSI)
I build custom platform integrations at the container runtime, storage, and networking layers. This includes implementing Cilium eBPF for secure network routing and observability, deploying custom container runtimes, and configuring CSI storage drivers on bare metal and public cloud.
Cilium eBPFCRI (Containerd)CSI Storage DriversTalos LinuxAPI Extensions
Pillar 02
Self Hosted AI Infrastructure & LLM Serving
Deploy secure, private AI runtimes, optimize hardware configurations, and run large language models on your own compute clusters.
Optimized LLM Inference Orchestration
I build low latency model serving infrastructures using dynamic runtimes. This includes setting up vLLM and TGI (Text Generation Inference) optimized for NVIDIA GPU clusters, as well as configuring llama.cpp for cost-efficient local execution on Apple Silicon.
vLLMllama.cppTGINVIDIA GPUsApple SiliconTensorRT LLM
Private AI Resource Management with LLMKube
Manage your private LLMs using the LLMKube operator. I automate model weight caching, configure multi tenant GPU sharing schedules, and implement autoscaling policies triggered by real time token throughput metrics.
LLMKube OperatorAutoscalingWeight CachingGPU SchedulingMulti Tenancy
Pillar 03
Stateful Kubernetes & AI Agent Gateways
Run databases declaratively on your own infrastructure, and connect agent networks using fast, secure routing layers.
Cloud Native Database Cluster Orchestration
I deploy stable transactional databases and messaging engines on Kubernetes. Using specialized operators like CloudNativePG, Vitess, and Strimzi, I automate streaming replication, point in time recovery (PITR), and zero downtime rolling upgrades.
CloudNativePGVitessStrimzi KafkaStatefulSetsPITR Backups
High Performance Agent Gateways & MCP Networks
I configure API proxy layers using the agentgateway architecture. This handles secure traffic routing across diverse Model Context Protocol (MCP) servers, implements token caching to reduce costs, and manages credentials safely.
agentgatewayModel Context Protocol (MCP)LLM RoutingToken CachingAgent Mesh
Ready to build high performance infrastructure?
Whether you need to extend Kubernetes via custom Go Operators, deploy HA transactional databases, or provision private AI serving clusters, let's build a stable, scalable foundation.