AI Systems Basics
What a language model does, and why the system around it decides whether it works.
Read in order
Series
Some ideas only land in sequence. Each series is a short run you can read start to finish, and every part also sits on the map next to the component it is about.
All 18 series complete, 100 parts.
60 Days of Production AI Systems
From LLM Demo to AI Product
The gap between a capable model and a product people can rely on: evidence, structure, and choosing the right shape.
Retrieval Foundations
Chunking, embeddings, vector stores, keyword and hybrid search, and the permissions retrieval must respect.
Evidence-Driven RAG Patterns
Reranking, provenance, multi-query and graph retrieval: turning search results into evidence.
Production RAG Operations
Knowledge in systems, multimodal evidence, freshness, recovery and evaluating retrieval in parts.
Agentic AI Foundations
Outcomes, loops, planning, tools, memory and reflection: what makes an agent more than a prompt.
Agent Control and Supervision
Human review, guardrails, recovery, budgets and when to split one agent into several.
Multi-Agent Design Patterns
Hierarchies, dispatch, pipelines, peers, shared state and debate, with the failure modes of each.
Agent Coordination Contracts
Roles, shared language, memory permissions, handoffs and decision rules between agents.
Production AI Readiness
Observability, evaluation, untrusted content, governance and infrastructure: what production asks of the whole system.
Five-day tracks
5 Days of MCP for Production AI
What the Model Context Protocol standardises, where servers run, and when MCP earns its overhead.
5 Days of Agent Infrastructure
The runtime underneath agents: autonomy, isolation, scaling and failure.
5 Days of Production RAG
Retrieval as a system you operate: trusted sources and freshness, ingestion as a data product, retrieval as its own subsystem, chunking for the question, and citations you can evaluate.
5 Days of AI Evaluation
Evaluation that can block a release: what an eval may stop, golden sets maintained like products, calibrated judges, regression gates and evaluation after launch.
5 Days of AI Safety Engineering
Safety as architecture rather than moderation: threat models, agent identity and scope, untrusted input, blast radius and data exfiltration.
5 Days of Multi-Agent System Design
When several agents actually help, and what it takes to run them: role contracts and handoffs, who controls the run, budgets and stop conditions, traces and recovery.
5 Days of AI Architecture Patterns
The building blocks of a production AI system: the model gateway, choosing the right pattern, control and data planes, approvals and tenant boundaries, and a reference architecture.
5 Days of AI Operations and Observability
Running AI in production: traces and spans, cost, latency and capacity, production evaluation signals, incidents and rollbacks, drift and an operating cadence.