LOCKEDIN / FORWARD-DEPLOYED ENGINEERING 101
Understand it.
Then build it.
You can be an experienced engineer and still be new to agents. Start with what each thing is and how it helps a person do useful work. Explore 48 original introductions across software, agents, cloud, security and delivery. Start here without a model key; each topic connects to deeper instruction and practice.
- 101UnderstandPlain language, visual walkthrough, vocabulary
- 102Build with guidanceOriginal lessons, examples and knowledge checks
- 103Practise and explainRunnable labs, failure cases and handover
- 01Retrieve within scope
- 02Select evidence
- 03Answer or abstain
A way to ground model responses
What is retrieval-augmented generation?
Retrieval-augmented generation (RAG) finds relevant information and supplies it to a model before it answers. It connects a model to a selected evidence collection without making that collection part of the model's learned parameters. Retrieval can use keyword search, semantic search or a combination.
Enterprise assistants need current company knowledge and clear boundaries on who may see it.
CONNECT IT TO WHAT YOU KNOW
Start with something familiar.
It is an open-book question: first find the right pages, then answer with those pages available. The quality of the answer depends on finding the right edition and reading it correctly.
RAG does not guarantee truth. The retrieved source can be wrong, stale, unauthorized or irrelevant, and the model can still misrepresent it.
Words worth knowing
- Retrieval
- Finding candidate information for a task.
- Grounding
- Connecting a response to supplied evidence.
- Abstention
- Choosing not to answer when support is insufficient.
ONE EXAMPLE / FICTIONAL CASE
Watch the work take shape.
Find the approved maintenance procedure for the caller's equipment, then draft a response citing that exact version.
- 01Retrieve within scope
- 02Select evidence
- 03Answer or abstain
Retrieve within scope
Filter by the caller's permissions before content enters the model context.
Authorized candidate passagesFROM RECOGNITION TO REASONING
Compare three possible search results.
- 01
Create a relevant current source, an outdated source and a source from another organization.
- 02
Decide which may enter the context.
- 03
Explain the response if only the latter two are available.
A source selection table with reasons.
Compare your reasoning when you are ready
Semantic similarity cannot override access or freshness. If no authorized current evidence supports the answer, return an explicit gap rather than synthesizing a guess.
Keep notes in your own working copy. Moving through these cards does not record completion.
TURN THE IDEA INTO USEFUL WORK
You can now go deeper.
RAG does not guarantee truth. The retrieved source can be wrong, stale, unauthorized or irrelevant, and the model can still misrepresent it.
Academy lessons use your existing training access. Free guides stay open.