The Opportunity

We are seeking a Platform Specialist (Director level) to serve as the organization's top technical authority on Kubernetes and the most senior hands-on engineer for CK-Kube, our Kubernetes Cost Intelligence platform. This is a deep individual contributor role — 60%+ hands-on engineering — where you will architect, implement, and technically lead CK-Kube as the principal engineer. You will set the technical direction, write production code, and drive architectural decisions. We are not looking for a people manager; we are looking for the strongest Kubernetes systems engineer we can find.

What You'll Own

CK-Tuner-Kubernetes — Kubernetes Cost Intelligence Platform

• Architect and implement the cost allocation engine — cluster, namespace, deployment, pod, and container granularity across EKS, AKS, and GKE
  

• Design and build the real-time data collection pipeline: agent architecture, ClickHouse time-series storage, gRPC streaming between agent and datastore
  

• Implement Karpenter integration for node lifecycle management and bin-packing optimization
  

• Build custom Kubernetes controllers and operators for cost policy enforcement and automated remediation
  

• Design shared cost distribution algorithms — system namespaces, control plane costs, networking overhead, idle capacity attribution
  

• Integrate CK-Tuner-Kubernetes with CK-Lens for a unified cloud + container cost view
  

Container Optimization Engine

• Design and implement container right-sizing algorithms for CPU and memory requests/limits based on real usage patterns
  

• Build node pool optimization logic — instance type selection, scaling policies, bin-packing efficiency scoring
  

• Implement Karpenter-based spot and preemptible node policies for fault-tolerant workloads
  

• Build the automated right-sizing execution pipeline via CK-Tuner integration
  

GPU Container Cost Intelligence

• GPU utilization tracking and idle GPU detection for AI/ML workloads running on Kubernetes
  

• Multi-cluster GPU cost comparison across EKS, AKS, and GKE
  

• Integration with the FinOps for AI initiative for GPU pod-level cost attribution
  

Responsibilities

Technical Leadership

• Serve as CK-Tuner-Kubernetes's principal architect and most senior hands-on engineer
  

• Set architectural standards and code quality bars; mentor engineers through technical pairing and design reviews
  

• Drive technical roadmap and architecture decisions in partnership with Product Management
  

Hands-On Engineering

• Write production Go code for CK-Tuner-Kubernetes's core systems: agent data collection, metrics processing, cost allocation engine
  

• Design and implement custom Kubernetes controllers and operators
  

• Build and optimize the ClickHouse time-series data model for cost metrics at scale
  

• Implement gRPC streaming with backpressure, circuit breakers, and mTLS between agent and datastore
  

• Develop Karpenter-based node optimization policies and consolidation algorithms
  

• Performance-tune the metrics pipeline: 10-second scrape intervals, 1-minute rollups, multi-cluster aggregation
  

Technical Strategy

• Design the agent data collection layer — hybrid metrics collection via Metrics API, Kubelet Summary, Kubelet Proxy, and optional Prometheus endpoints
  

• Architect the ClickHouse time-series schema with materialized views for multi-resolution aggregation (5m, 1h, 1d)
  

• Build the delta processing pipeline — in-memory state comparison with ring buffers (discovery 10K, metrics 50K, events 100K)
  

• Design cost allocation algorithms for shared resources — control plane, networking, system namespaces, idle capacity
  

• Architect multi-cloud Kubernetes support (EKS primary, AKS/GKE Phase 4) with provider-specific pricing API integrations
  

• Build integration points with CK-Lens, CK-Tuner, and CK-Intelligence
  

Technical Landscape You'll Navigate

Kubernetes & Container Orchestration

• Platforms: EKS (Fargate, managed node groups), AKS, GKE (Autopilot, standard), on-prem Kubernetes
  

• Ecosystem: OpenCost, Karpenter, Helm, Kubernetes Operators, K8s API Server
  

• Resource Management: Requests/limits, node autoscaling, pod scheduling, bin-packing, spot/preemptible nodes
  

• Kubernetes Internals: Custom controllers, operators, CRDs, admission webhooks, scheduler plugins, informers, leader election, reconciliation loops
  

Data Engineering

• ClickHouse (time-series analytics), Apache Pulsar/NATS JetStream (message broker), gRPC bidirectional streaming with backpressure
  

Cloud Providers

• AWS: EKS, Fargate, EC2 (GPU instances), S3, CloudWatch, Cost & Usage Reports
  

• Azure: AKS, Azure Monitor, Azure Billing APIs
  

• GCP: GKE, GKE Autopilot, BigQuery Billing Export
  

Requirements