24 ott - Italia
Mercor
h3About the job /h3 /brpMercor connects elite creative and technical talent with leading AI research labs. Headquartered in San Francisco, our investors include Benchmark, General Catalyst, Peter Thiel, Adam D'Angelo, Larry Summers, and Jack Dorsey. /p /brpPosition: PDF Annotation Transcription Experts – Korean Type: Contract Compensation: $34/hour
Location:
Remote /p /brh3Role Responsibilities /h3 /brul /brliSource documents by finding publicly accessible Korean PDFs with multimodal elements like images, tables, or diagrams. Record the source of each document. /li /brliAnnotate structure by identifying and bounding every meaningful region of the page. Assign each a component type and a reading-order index. /li /brliRecord relationships by linking each region to its parent component, such as a figure or table. /li /brliTranscribe faithfully by reproducing all text exactly as it appears in Hangul, including any hanja and handwritten content. Flag any illegible regions. /li /brliCapture page metadata, including language, document type, source, page dimensions, and flags for tables, formulas, and handwriting. /li /brliReview a colleague's work.
Every task is reviewed end-to-end by a second Korean expert. /li /br /ul /brh3Qualifications Must-Have /h3 /brul /brliNative Korean speaker with full command of Hangul, including hanja. /li /brliExperience in bilingual transcription, translation, editorial work, or AI training data. /li /brliExact character-level accuracy. /li /brliSystematic application of taxonomy across pages. /li /brliComfort with unfamiliar layouts like multi-column newspapers and handwritten forms. /li /br /ul /brh3Preferred /h3 /brul /brliExperience with AI training data : annotation, labeling, grading, or bilingual evaluation. /li /brliSkills in transcription and localization: MTPE, subtitling, bilingual QA, OCR correction. /li /brliDocument production
experience:
typesetting, copy-editing, proofreading, or digitization of Korean-language material. /li /brliKnowledge of script and encoding: Unicode normalization, Korean input methods, Hangul jamo composition, and hanja handling. /li /br /ul /brh3Application Process /h3 /brp(Takes 20–30 mins to complete) /p #J-18808-Ljbffr
24 ott - Italia
Etjca Group
24 ott - Bergamo
Randstad Italia
24 ott - Trezzano sul Naviglio
MYA TECH
24 ott - Alto
MYA TECH