Notes from the engineering desk.
Opinionated, evidence-led writing on the systems we build — AI engineering, cloud and FinOps, modernization, and architecture.
Beyond Exit Code 137: Architecting K8s Memory with VPA, Karpenter, and QoS
Prevent Kubernetes OOMKills using QoS classes, VPA, Karpenter, in-place resizing, and memory-efficient application architectures.
Read more
How MCP Changes Cloud-Native AI Architecture
Model Context Protocol (MCP) reshapes cloud-native AI architecture with tool federation, MCP server topology, cloud security, and multi-cloud systems.
Read more
Why Kubernetes Is Being Adopted Widely Today?
Discover why Kubernetes has become the leading container orchestration platform, enabling scalability, resilience, automation, and cloud-native innovation.
Read more
Kubernetes ETL Pipelines for ML: MLOps Best Practices & Production Guide
Build scalable ETL pipelines for ML on Kubernetes. Master K8s Jobs, Argo Workflows, GPU scheduling, security, and MLOps production best practices.
Read more
Confidential AI on Kubernetes: Secure LLM Inference
Learn how Confidential AI secures LLM inference on Kubernetes by protecting prompts, model weights, RAG context, and runtime data during active processing.
Read more
Running Production RAG Workloads on Kubernetes
Running Production RAG Workloads on Kubernetes: architecture, autoscaling, observability, security, and performance optimization strategies.
Read more
Designing Zero-Downtime Database Migrations in Kubernetes
Learn how to perform zero-downtime database migrations in Kubernetes using the Expand-Contract pattern, migration jobs, and backward-compatible schema changes.
Read more
Kubernetes Secrets Management in 2026: Best Practices
Kubernetes Secrets Management in 2026: production best practices for encryption at rest, RBAC, rotation, ESO, Vault, and preventing secret leaks.
Read more
Explore Workload-Aware Scheduling with PodGroups in Kubernetes
Discover how Workload-Aware Scheduling (WAS) and PodGroups in Kubernetes prevent deadlocks, optimize resource allocation, and scale AI/ML training workloads.
Read more
When Karpenter Saves Money in Amazon EKS And When It Doesn't
Discover when Karpenter delivers real AWS cost savings over EKS Managed Node Groups, including ROI, Spot optimization, and migration payback.
Read more
Part 2 Beyond the Exit Code: K8s Runtime Traps
Diagnose Kubernetes OOMKilled events beyond cgroup metrics. Deep-dive into Go RSS inflation, JVM native leaks, eBPF allocation tracing, and page cache pitfalls.
Read more
How Internal Developer Platforms Simplify Kubernetes for Developers ?
Learn how Internal Developer Platforms (IDPs) simplify Kubernetes by abstracting complexity, enabling self-service deployments, and improving developer productivity.
Read more