AI Edge Course: Agents, Ontologies & Keeping Your AI on Track

Ontologies, schemas, and the machinery that stops an agent from confidently doing the wrong thing.

Self-paced  ·  Available now - scroll for details  ·  drC / The AI Edge

An embedding tells you which text is closest. It cannot tell you which text is correct. Retrieval inherits that blindness — it will hand your agent the topically nearest chunk with complete confidence, including the chunk that says the opposite of what's true.

Prompting doesn't fix this. You can write "only use the retrieved context" in capital letters eight times and it will work most of the time, which in an agentic system — where one bad output becomes the next step's input — is another way of saying it fails.

This course is about the other kind of instruction: the kind that guarantees rather than guides. Schemas the API enforces. Validators that reject bad data before it moves. And at the top of the stack, a formal model of your domain that a reasoner checks every proposed action against — so this refund exceeds the original charge gets caught by machinery, not by hoping the model noticed.

The guardrail stack

Four layers. Each catches what the one above it can't.

Prompt
Nothing, reliably. It guides.
System prompts, few-shot examples
Schema
Malformed output — wrong shape, missing fields
tool_use with JSON Schema
Validation
Semantically wrong values — out of range, wrong type, broken business rule
Pydantic, retry loops
Ontology
Statements that are well-formed, plausible, and impossible in your domain
RDFS, OWL, SHACL, a reasoner

Most courses stop at layer three. Layer four is where the interesting failures live.

The curriculum

Click on the sections below to see details of each.
drC · The AI Edge

Five layers, one stack

Each layer answers a different question. Go top to bottom, or jump to the one you need.
Click any layer to see its lessons. Click again to close.

Ready to start?

Click the link to get started. At completion you’ll exit with a portfolio that proves you can build production AI systems — and get a sense of what Anthropic thinks is important by exploring their Claude Certified Architect (CCA-F) certification exam.