The measured state of the job index.

>=1,000 measured Gold-job apply links, >=98% live, measured within 8 days

Measured value

Not measured

No retained apply-link liveness artifact was found.

Every described default-served job is counted exactly once as true/false/null for both auth fields with current extractor provenance; retained owner-signed >=200-row review has >=95% emitted-value precision and zero known negation inversions

All described jobs reported

Yes

Current extractor provenance rate

99.3%

Current extractor provenance rows

1,331,058

Described jobs

1,340,151

Default-served jobs with descriptions plus the retained auth-flags calibration. Null means the source is silent; it is never coerced to false.

Change outbox holds <= 5,000,000 rows

Estimated rows

0

Max rows

5,000,000

Measured by jobs_supply.data_quality.

100% of companies with open jobs have a syntactically valid domain

Coverage rate

77.8%

Total

84,927

Valid domains

66,047

Measured by jobs_supply.data_quality.

500 unique human-reviewed rows; publish measured baseline precision without substitution

Artifact sha256

Not measured

Blocked reason

human-reviewed calibration artifact not found

Precision

Not measured

Reviewed

0

Retained jobs-supply accuracy calibration artifact.

100% human-readable and no opaque company card

Coverage rate

99.9%

Human readable

84,856

Opaque companies with resolvable jobs

16

Total

84,927

Measured by jobs_supply.data_quality.

Zero tal.af or Jobs in the Wild references

Measured value

0

Measured by jobs_supply.data_quality.

Top-country stats count equals country-search total

Measured by jobs_supply.data_quality.

500 unique human-reviewed rows; publish measured baseline precision without substitution

Artifact sha256

Not measured

Blocked reason

human-reviewed calibration artifact not found

Precision

Not measured

Reviewed

0

Retained jobs-supply accuracy calibration artifact.

>=500 reviewed jobs across >=500 human-labeled pair cases; false merges <2%, missed duplicates <10%

Candidate missed duplicate pairs

0

False merge rate

Not measured

Gold set case count

500

Gold set size

0

Measured from the committed human-adjudicated dedup gold set against current jobs-db group ids.

>=40% of described default-served jobs have skills and a retained >=200-job review measures >=95% emitted-label precision for both skills and benefits

Benefits coverage rate

6.2%

Benefits mean items per described job

0.1

Default served open jobs

1,340,151

Described jobs

1,340,151

Default-served jobs with source descriptions plus the retained human-reviewed derived-label artifact.

Zero non-canonical employment_type, work arrangement, source, or status values

Invalid employment type

0

Invalid source

1,150

Invalid status

0

Invalid work arrangement

1

Measured by jobs_supply.data_quality.

No missing seen dates or open rows with closed_at

Missing first seen

0

Missing last seen

0

Open with closed at

0

Measured by jobs_supply.data_quality.

No description backlog, or a hydration wave ran within 3.0h

Hydration age hours

0.11

Hydration backlog open

1,330,439

Last hydration run at

Oct 17, 2026, 5:27 AM UTC

Max age hours

3

Measured by jobs_supply.data_quality.

Backlog under 50k, or >=10,000 rows hydrated in 24h

Hydrated 24h

100,126

Hydration backlog open

1,330,439

Min rows 24h

10,000

Measured by jobs_supply.data_quality.

A board load landed within the last 30.0h

Last sweep at

Oct 17, 2026, 2:52 AM UTC

Max age hours

30

Sweep age hours

2.68

Measured by jobs_supply.data_quality.

>=90% of day-old intake clears full job and company Gold

Cohort first seen 24 48h

431,729

Cohort gold

128,134

Gold rate

29.7%

Measurement

current full Gold share of retained unique intake aged 24-48h; not a historical 24h deadline measurement

Measured by jobs_supply.data_quality.

Non-board Gold jobs require location; board location gaps remain measured but do not block publication

Legacy country array mismatches

86,946

Legacy derived country missing from codes

0

Public gold jobs

2,096,936

Public gold location measured

Yes

Measured by jobs_supply.data_quality.

100% of publicly served jobs satisfy age, description, freshness, and human-link rules; non-board jobs also require location

Default served jobs

1,340,151

Described gold jobs

1,340,151

Excluded dated over 6 months

396,057

Excluded description below 200

1,330,439

Measured by jobs_supply.data_quality.

No stale-open backlog, or a closure applied within 36.0h

Closure age hours

0.34

Last applied closure at

Oct 17, 2026, 5:13 AM UTC

Max age hours

36

Stale open 30d

753,196

Measured by jobs_supply.data_quality.

Every explicit source-salary statement is retained or parsed; parsed ranges are valid

Malformed currency

0

Negative or inverted

0

Raw present

365,083

Salary populated

365,083

Measured by jobs_supply.data_quality.

>=95% of otherwise default-served open jobs verified on-board within 21 days

Coverage rate

78.4%

Eligible open jobs

4,030,520

Freshness lag p50 days

5.17

Freshness lag p95 days

50.48

Measured by jobs_supply.data_quality.

500 unique human-reviewed rows; publish measured baseline precision without substitution

Artifact sha256

Not measured

Blocked reason

human-reviewed calibration artifact not found

Precision

Not measured

Reviewed

0

Retained jobs-supply accuracy calibration artifact.

Every open job has at least one complete field-level provenance envelope

Coverage rate

64.1%

Jobs with complete envelope

2,651,845

Open jobs

4,136,529

Measured by jobs_supply.data_quality.

100% canonical family, subfamily, and skill-cluster shape; honest unknown is canonical

Missing skill clusters

0

Missing subfamily

198,154

Noncanonical family

1,099,466

Null or empty family

1,430

Measured by jobs_supply.data_quality.

500 unique human-reviewed rows; publish measured baseline precision without substitution

Artifact sha256

Not measured

Blocked reason

human-reviewed calibration artifact not found

Precision

Not measured

Reviewed

0

Retained jobs-supply accuracy calibration artifact.