Issues

fix(visibility): roster adapter starves working repair lanes of lead profile data

#2154

· quntao-z opened

on Aug 28, 2026

fix(scrapers): materializer writes bypass all schema enums (no runValidators)

#2137

· quntao-z opened

on Aug 28, 2026

fix(model): Observation.entityType enum contradicts 72% of the corpus

#2135

· quntao-z opened

on Aug 28, 2026

refactor(dedupe): converge on one before-mint resolver + one dedup engine

#2063

· quntao-z opened

on Aug 27, 2026

chore(scrapers): retire acceptingUndergrads binary + legacy-v1 prompt path end-to-end

#2055

· quntao-z opened

on Aug 27, 2026

scrapers: coverage backlog and skip decisions (migrated from registries)

#2040

· quntao-z opened

on Aug 26, 2026

feat(scrapers): add development-resume sweep mode to pick up after an interrupted full sweep

#2006

· quntao-z opened

on Aug 26, 2026

refactor(pipeline): make materialize->gate->index idempotent to retire the fix* repair genre

#1945

· quntao-z opened

on Aug 25, 2026

data-quality(descriptions): FRA description harvest misses a cross-appointed faculty member's secondary official Yale profile (e.g. medicine.yale.edu) when the primary department profile is a bare stub

root-cause

#1937

· quntao-z opened

on Aug 25, 2026

data-quality(descriptions): faculty bio-opener repair blanks whole fullDescription when every sentence is name/pronoun-led, stranding research-rich FRAs

#1936

· quntao-z opened

on Aug 25, 2026

data-quality(identity): surname-collision medicine.yale.edu profile bound as a real faculty member's PRIMARY_IDENTITY (wrong url + fabricated researchArea + latent profile_identity_risk)

root-cause

#1935

· quntao-z opened

on Aug 25, 2026

data-quality(users): shared cross-field topics cluster mis-attributed onto users directory records (upstream of fabricated entity researchAreas)

root-cause

#1934

· quntao-z opened

on Aug 25, 2026

Navigation Menu

Uh oh!

All issues

fix(visibility): roster adapter starves working repair lanes of lead profile data

fix(scrapers): materializer writes bypass all schema enums (no runValidators)

fix(model): Observation.entityType enum contradicts 72% of the corpus

refactor(dedupe): converge on one before-mint resolver + one dedup engine

chore(scrapers): retire acceptingUndergrads binary + legacy-v1 prompt path end-to-end

scrapers: coverage backlog and skip decisions (migrated from registries)

feat(scrapers): add development-resume sweep mode to pick up after an interrupted full sweep

refactor(pipeline): make materialize->gate->index idempotent to retire the fix* repair genre

data-quality(descriptions): FRA description harvest misses a cross-appointed faculty member's secondary official Yale profile (e.g. medicine.yale.edu) when the primary department profile is a bare stub

data-quality(descriptions): faculty bio-opener repair blanks whole fullDescription when every sentence is name/pronoun-led, stranding research-rich FRAs

data-quality(identity): surname-collision medicine.yale.edu profile bound as a real faculty member's PRIMARY_IDENTITY (wrong url + fabricated researchArea + latent profile_identity_risk)

data-quality(users): shared cross-field topics cluster mis-attributed onto users directory records (upstream of fabricated entity researchAreas)

Uh oh!

Search results