AINA Data Engine Room · title taxonomy · 2026-06-15

PE Donor Title Taxonomy Gate

Deterministic donor buckets preserved without row promotion.

The Single Idea

The donor buckets are useful lineage, but the audit is unscored. Current engine-room title, JD, and source-authority receipts stay serving truth.

2026donor bucket rows
50audit template rows
0row, runtime, or embedding authority

Mapped Surfaces

SurfaceValidMetrics
beta_readiness_path_v1True{"coverage_title_rows": 74225, "reviewed_residual_hold_count": 491, "serviceable_title_rows": 50053}
production_source_authority_registry_v1True{"clean_candidate_row_count": 44440, "jobs_research_title_audit_rows": 37478, "trusted_jobs_research_titles": 15104}
jd_aware_role_context_evidence_v1True{"role_context_row_count": 1056, "rows_with_jd_summary": 1014, "top_1000_titles_with_role_context": 964, "top_500_titles_with_role_context": 484}
top_worked_title_readiness_v1True{"icp_serviceable_top_title_count": 1000, "production_unlock_count": 0, "semantic_sample_row_count": 50}
source_authority_registry_v2True{"registry_row_count": 40, "vector_count": 151983}

Checks

CheckStatus
current_engine_room_replacement_proof_presentPASS
current_engine_room_title_receipts_validPASS
donor_audit_template_existsPASS
donor_audit_template_has_50_rowsPASS
donor_audit_template_is_unscoredPASS
donor_bucket_count_matches_provenancePASS
donor_bucket_jsonl_existsPASS
donor_output_is_deterministic_no_llmPASS
donor_provenance_existsPASS
donor_repos_read_onlyPASS
jd_context_receipt_present_for_top_bandPASS
labels_are_metadata_not_truthPASS
legacy_reviewer_fields_absentPASS
no_live_gemini_api_invokedPASS
no_public_runtime_or_external_writesPASS
no_runtime_embedding_or_batch_authorityPASS
package_validPASS
row_promotion_blocked_until_audit_or_replacement_diffPASS
source_refs_are_hash_or_external_refs_onlyPASS
title_taxonomy_source_row_presentPASS