30-day run history
Daily pipeline activity
New postings New assessments
System primer
How the daily pipeline works
The system finds company job boards, extracts recent US openings, then runs two independent branches: a broad public board and candidate-specific fit scoring.
How to read the numbers: “This run” is activity from the latest completed stage run; “current total” is the latest stored or published inventory; rolling metrics state their time window.
Independent review
Gemma usage
Current period—
01
Latest discovery run
Job Discovery
View source and board diagnosticsDiscovery sources · new additions
Discovery sources
New board additions
02
Latest extraction run
Job Extraction
View ATS reliability diagnosticsFailure rate · error taxonomy
Board reliability by ATS
| ATS family | Boards selected | Boards failed | Failure rate | Signal |
|---|
Error taxonomy
03
Latest public-board build
Public Discovery Board
View enrichment diagnosticsFive model passes · cache · descriptions
Five-pass enrichment
Cache and description intake
04
Latest market-analysis run
Market Analysis
View Spark pipeline diagnosticsStage timings · data quality · deduplication
Spark stage timings
Aggregate output
05
Latest assessment run
Relevancy Assessment
View model and sync diagnosticsRetries · fallbacks · tokens · calls
Assessment reliability
| Step | Attempted | Failed | Fallback | Error rate |
|---|