Retrieval and embeddings · Lesson 15 of 16 · 5 min

Build retrieval evidence

A retrieval evidence bundle records query, filters, index and embedding revisions, selected source IDs and failure cases. Redact sensitive query content when sharing outside its boundary.

Open the lab Course map

Mastery 0/3Run itExplain itApply it

learner@clairos:~ $

↑ history · Tab complete · Ctrl+C cancel · Esc, Tab to leaveExit 0

Your objectives

Hints are always available. Wrong turns never cost XP.

1
Read the case evidence
Read /opt/ai-lab/cases/aiops-retrieval-embeddings-015.json. Separate the recorded facts from your assumptions.
Hint & worked example
A retrieval evidence bundle records query, filters, index and embedding revisions, selected source IDs and failure cases. Redact sensitive query content when sharing outside its boundary.
The complete case JSON is read successfully.
```
cat /opt/ai-lab/cases/aiops-retrieval-embeddings-015.json
```
Common mistake to avoid: Source IDs alone are enough to reproduce every query.
2
Isolate the deciding signal
Select the filters_recorded field from the case JSON. What does it establish—and what does it not establish?
Hint & worked example
Add the applied filters so another operator can reproduce the retrieval result.
A successful JSON query isolates filters_recorded.
```
jq .filters_recorded /opt/ai-lab/cases/aiops-retrieval-embeddings-015.json
```
Common mistake to avoid: Using a different case or an unrelated signal as the justification.
3
Record your operational decision
Choose the justified code: record-query-filters, change-without-evidence, or ignore-the-signal. Save that one code followed by a newline in /home/learner/lab/aiops-retrieval-embeddings-015-decision.txt.
Hint & worked example
Add the applied filters so another operator can reproduce the retrieval result.
The decision artifact contains the correct code, not just a matching command in history.
```
printf '%s\n' 'record-query-filters' > /home/learner/lab/aiops-retrieval-embeddings-015-decision.txt
```
Common mistake to avoid: Source IDs alone are enough to reproduce every query.

Understand the pattern

A retrieval evidence bundle records query, filters, index and embedding revisions, selected source IDs and failure cases. Redact sensitive query content when sharing outside its boundary.

Interpret the evidence

Add the applied filters so another operator can reproduce the retrieval result.

A tempting mistake

Source IDs alone are enough to reproduce every query.

All names, measurements, prices, versions, and thresholds in this exercise are fictional training data. The evidence supports this bounded scenario, not a claim about a live provider.

Worked example and expected result

$ cat /opt/ai-lab/cases/aiops-retrieval-embeddings-015.json

{
  "fixture": "Fictional training data; not a live measurement.",
  "index": "index-42",
  "embedding": "embed-a-v2",
  "filters_recorded": false,
  "source_ids_recorded": true
}

Exit status: 0. The command completed successfully in the training fixture.

$ jq .filters_recorded /opt/ai-lab/cases/aiops-retrieval-embeddings-015.json

false

Exit status: 0. The command completed successfully in the training fixture.

$ printf '%s\n' 'record-query-filters' > /home/learner/lab/aiops-retrieval-embeddings-015-decision.txt

(no output; exit 0)

Exit status: 0. The command completed successfully in the training fixture.

Try a variation

Name the assumption that would have to change for “record-query-filters” to stop being the justified action. Do not silently replace missing evidence with a guess.

Explain it

What does the evidence in “Build retrieval evidence” justify?Add the applied filters so another operator can reproduce the retrieval result.Source IDs alone are enough to reproduce every query.Make the change now and gather the missing evidence afterward.

Apply it elsewhere

A colleague concludes: “Source IDs alone are enough to reproduce every query.” What should you do?Accept it because the training command completed.Treat this fixture as a live measurement.Reject that conclusion; return to the case evidence and its stated limits.

Keep the momentum.

Complete the lab, then explain and apply the pattern.

Next: Retrieval checkpoint Try an incident drill

Anonymous lesson and mastery progress lasts for this session. Sign in to keep synced lesson XP between sessions; mastery stays device-local. No speed bonus, lost lives, or hint penalties.

Keep learning

Case methodology: evidence, limits, decision