A cited retrieval agent using an approved public document set and a ten-question evaluation harness.
BRAINS · Masterclass
Retrieval and Working Memory
Retrieve evidence without turning everything into permanent context.
Design source authority, ingestion, retrieval, evaluation, freshness, and evidence-limited responses.
Evaluation table, two explicit evidence-limited responses, one corrected retrieval failure, and a deletion test.
Cited Retrieval Lab
Class requirement: create or sign in to your Alternate Clouds account, deploy the Cited Retrieval Lab template, and submit its live URL before the class is marked complete.
Create your Alternate Clouds account ↗Use only work you own or are authorized to use.
Every submission needs a source-and-license manifest covering text, code, images, audio, video, fonts, datasets, models, weights, and AI reference inputs. Attribution alone is not permission. Questionable material stays private or is replaced before deployment or showcase.
Ask about rights clearance01Memory is not one thing+
Separate context windows, session state, scratchpads, retrieval, and source-of-truth data.
Map each memory type in your project and identify who can correct or delete it.
02From source to index+
Use source authority, structure-aware chunks, metadata, embeddings, lexical search, and reranking deliberately.
Create an ingestion map with ownership, version, date, permissions, and deletion behavior.
03Retrieval evaluation+
Measure recall, precision, citation correctness, faithfulness, freshness, latency, and cost separately.
Write ten answerable and unanswerable questions before tuning retrieval.
04Memory operations+
Plan re-indexing, contradiction handling, deletion propagation, and changing truth over time.
Insert one stale document and prove the current source wins.
The complete course adds the teaching deck, Alternate Clouds lab, source-bounded guide, review rubric, and instructor feedback.
Find a class