GCP Spotlight
Unified FinOps platform that analyzes cloud resource efficiency across an entire GCP organization — covering Kubernetes clusters, virtual machines, cloud storage, and AI workloads in a single pane of glass. Four service modules: GKE (59 clusters, adaptive right-sizing with 25–45% headroom by workload variability), GCE (2,222 VMs, idle detection, CUD coverage analysis), GCS (2,989 buckets, lifecycle optimization, egress monitoring), and AI/Vertex AI (model spend, anomaly detection, chargeback by project).
Full-stack build — Python FastAPI backend (123 API endpoints), React/TypeScript/Tailwind frontend (12 pages, 61 components), three isolated SQLite databases with WAL mode, and a hybrid cost model that layers BigQuery billing data, derived per-unit rates, and on-demand fallback pricing. Deployed to GKE via ArgoCD with Workload Identity and Terraform IAM across ~600 monitored projects.