Kubernetes Autoscaling & Capacity Optimization

Overview

Built and maintained AWS EKS node provisioning with Karpenter by defining scalable, workload-aware node pools to automatically right-size capacity and improve cluster elasticity for data/compute services.

Achievements

Enabled faster scale-out and more cost-efficient Kubernetes compute by introducing Karpenter pools tuned for different workload profiles, while reducing EC2 spend via right-sizing and lifecycle optimization.

Responsibilities

• Designed Karpenter node pools and provisioning constraints to match workload requirements (CPU/memory, architecture, zones, taints/tolerations), improving scheduling reliability and reducing manual node management.
• Automated cluster capacity management by codifying pool policies (limits, consolidation/expiration, disruption controls) to support safe scaling and predictable operations.
• Optimized EC2 costs by selecting appropriate instance families/sizes, leveraging spot where suitable, and eliminating over-provisioned capacity through right-sizing and consolidation strategies.
• Standardized environment configuration for repeatable deployments across stages (e.g., QA/prod) through infrastructure-as-code practices.