▸ Blog
Read it as a series.
Related posts, ordered into a path. Start at part one and work through — or jump to whatever you need.
- 6 parts
RAG in Production
Everything that breaks when retrieval-augmented generation meets real users — and how to fix it.
- 01 Why your RAG implementation is failing in production
- 02 7 mistakes wrecking your production RAG stack
- 03 Picking the right RAG stack: vector databases for AI
- + 3 more →
- 4 parts
Self-Hosting SaaS with Coolify
Ship and scale a SaaS on your own infrastructure with Coolify and Docker — without a platform bill.
- 01 Coolify in 2026: 60k GitHub stars and the self-hosted PaaS bet
- 02 Ditch expensive cloud providers for self-hosted SaaS
- 03 Self-hosted SaaS with Coolify and Docker
- + 1 more →
- 6 parts
Shopify Plus, Deep
Themes, headless, web components, and agentic commerce — building serious storefronts and apps on Shopify Plus.
- 01 Shopify Liquid vs. headless: picking the right stack
- 02 Shopify Storefront Web Components: headless light
- 03 Shopify UCP quick-start: make your store agent-ready
- + 3 more →
- 12 parts
AI Architecture Decisions
The architectural forks every AI product hits — ML vs GenAI, RAG vs fine-tuning, context vs memory, vectors vs graphs, and where MCP fits.
- 01 Machine learning vs generative AI
- 02 AI vs machine learning: an engineering deep dive
- 03 LLM vs AI agent: from prompts to action
- + 9 more →
- 11 parts
Infrastructure Decisions
Load balancers, proxies, gateways, containers, and scaling — the infrastructure trade-offs behind a system that stays up.
- 01 Horizontal vs vertical scaling
- 02 Load balancer vs reverse proxy: scale vs security
- 03 Load balancer vs API gateway
- + 8 more →
- 4 parts
Async & Messaging
Queues, brokers, and event-driven patterns — how to move the heavy work off the request path without losing data.
- 01 Synchronous vs asynchronous communication
- 02 Message queues for heavy-duty document processing
- 03 Scaling with RabbitMQ: why message brokers matter
- + 1 more →
- 8 parts
From Vibe Coding to Agentic Engineering
How AI-assisted development grows up — from conversational vibes, to MCP-driven agentic loops, to the review and verification gates that let it ship to production.
- 01 Vibe coding: why projects need more than just logic
- 02 AI vs traditional development: which fits?
- 03 Vibe coding: the shift to agentic workflows
- + 5 more →
- 6 parts
Engineering Roles, Compared
DevOps vs MLOps, SRE vs platform, cloud vs DevOps, data engineer vs scientist — who does what, and which role you actually need.
- 01 DevOps vs MLOps: key technical differences for 2026
- 02 SRE vs Platform Engineer: who to hire for scale
- 03 Cloud engineer vs DevOps engineer in 2026
- + 3 more →
- 3 parts
Monolith and Modularity
Choose an architecture, keep the modules honest, and only split when the friction earns it — the full monolith-to-services arc.
- 01 Monolith vs microservices: how to choose
- 02 Modular monoliths first: why microservices can wait
- 03 Monolith to microservices: a pragmatic guide
- 3 parts
Serving AI in Production
GPU-aware routing, scaling triggers that actually track LLM load, and the rate limits that stop one agent loop from burning your margin.
- 01 GPU-aware load balancing for AI inference
- 02 Smart auto-scaling for modern AI apps
- 03 Rate limiting: protecting your AI wallet
- 3 parts
Production Operations
Ship it, provision it, and see inside it — the CI/CD, infrastructure-as-code, and observability disciplines behind a system you can debug at 3am.
- 01 CI vs CD: automating quality and delivery
- 02 Terraform vs Ansible: the real difference in DevOps
- 03 Logging vs monitoring: a guide for scaling
- 3 parts
Performance and Limits
Diagnose the real bottleneck, cut the distance with edge delivery and caching, then cap the traffic before it caps you.
- 01 Bandwidth vs throughput: a wider pipe won't fix lag
- 02 CDN vs cache: why your high-traffic site needs both
- 03 Rate limiting vs throttling
- 6 parts
Laravel at Scale
Multi-tenant isolation, Octane workers that stay warm, and the replication-plus-backup setup that survives both hardware failure and human error.
- 01 Laravel multi-tenancy: a scalable SaaS architecture
- 02 Laravel Octane for high-traffic PHP apps
- 03 Replication vs backup: why Laravel needs both
- + 3 more →