quyennv.com

Senior DevOps Engineer. Notes on cloud, Kubernetes, and reliable systems — written clearly, kept for reference.

More notes

View all

LLM Architecture: Production Inference, RAG, Scaling, and Infrastructure Design

End-to-end LLM architecture for production: inference pipeline, model serving, RAG, fine-tuning trade-offs, GPU scaling, routing, guardrails, observability, and deployment patterns for AI infrastructure teams.

CAP Theorem in System Design: Consistency, Availability, and Partition Tolerance

CAP theorem explained for distributed systems: the three guarantees (C, A, P), trade-offs (CA, CP, AP), partition examples, real-world use cases, and diagrams.

Self-Hosted CI/CD Agent with AKS and Azure DevOps – Part 3: Kaniko in Detail

How Kaniko builds container images on Kubernetes without Docker. Job manifest, build context, ACR push, and pipeline integration with Azure DevOps.

Istio Service Mesh: Traffic, Resiliency, Observability, and Security

Introduction to Istio: control plane and data plane, sidecar vs ambient mesh modes (with examples), traffic control, resiliency, observability, and security.

Context Engineering for AI Agents: Summary and Deep Dive

A detailed, practical guide to context engineering for AI agents: context budgets, prompt vs context engineering, tool/context curation, just-in-time retrieval, and long-horizon strategies (compaction, structured notes, multi-agent).

Building a Lightweight Windows System Monitor Widget in C++

Step-by-step walkthrough of the SysMonitor widget: architecture, windowing model, metric collectors, rendering pipeline, tray integration, and ideas for extending it.

Designing Data‑Intensive Applications – Detailed Summary and Notes

Easy‑to‑understand summary of Martin Kleppmann’s Designing Data‑Intensive Applications (DDIA): reliability, scalability, maintainability, data models, storage, distributed systems, batch and stream processing.

Kubernetes PV, PVC, and How They Attach to Pods (Deployments, StatefulSets, DaemonSets)

PersistentVolume vs PersistentVolumeClaim, StorageClass, binding lifecycle, access modes, and how each workload type (Deployment, StatefulSet, DaemonSet) uses volumes with Pods.

Thinking in Systems: A Primer

Key ideas from Donella Meadows' Thinking in Systems—stocks and flows, feedback loops, the systems zoo, why systems surprise us, system traps, leverage points, and living in a world of systems.