Your guide to building
intelligent systems.
A free weekly publication for engineers who build AI-powered systems. No hype, no fluff — just architecture decisions, production patterns, and hard-won engineering lessons.
Built for engineers, not tourists.
Architecture deep-dives
RAG pipelines, multi-agent patterns, vector databases, and production infrastructure.
Production engineering
SDK integration, API design, testing strategies, and real-world .NET + Azure patterns.
Honest experiments
Coding agents, AI code-review, knowledge-base prototypes — tested and documented.
Model evaluations
Claude vs GPT, fine-tune vs RAG, small models, token economics — beyond vibes.
The four tracks.
AI Architecture
System design patterns for AI-powered applications. RAG pipelines, orchestration layers, and infrastructure decisions.
6 articles →Engineering
Practical engineering for AI systems. SDK integration, API design, testing strategies, and production-grade patterns.
5 articles →Models
LLM evaluation, fine-tuning strategies, model selection, benchmarking, and cost-performance analysis.
5 articles →Experiments
Hands-on explorations with AI tools, agent frameworks, coding assistants, and emerging techniques.
5 articles →Four steps to get started.
Subscribe (free)
Get every new article in your inbox. No paywall, no spam.
Pick a track
Choose AI Architecture, Engineering, Models, or Experiments.
Explore the knowledge tree
Each track has an interactive map showing how topics connect.
Go deep
Read full articles with diagrams, code samples, and production notes.
Popular articles to start with.
Designing RAG Systems That Actually Scale
The most important AI architecture pattern in production today.
Integrating Azure OpenAI with ASP.NET Core
Connect LLMs to real enterprise systems safely.
ModelsClaude vs GPT for Engineering Workflows
Practical model comparison beyond benchmarks.
ExperimentsTesting Autonomous Coding Agents
See how AI tools perform in real engineering workflows.
AI ArchitectureMulti-Agent Architecture Patterns
Orchestrator, supervisor, and swarm patterns with real trade-offs.
EngineeringPrompt Engineering as Software Engineering
Treat prompts like code — version, test, and iterate.