Mercor 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.

Source documents by finding publicly accessible Japanese PDFs with multimodal elements and record their origins.

Annotate structure by identifying and bounding every meaningful region, assigning component types and reading-order indices.

Record relationships by linking regions to their parent components like figures or tables.

Transcribe faithfully, reproducing all text exactly as it appears, including kanji, hiragana, katakana, and handwritten content.

Capture page metadata such as language, document type, source, page dimensions, and flags for tables and handwriting.

Review a colleague's work, ensuring tasks are reviewed end-to-end by a second Japanese expert.

Professional experience in interpretation, journalism, transcription, translation, or editorial work.

Exact character-level accuracy, prioritizing precision over speed.

Systematic approach, applying a taxonomy consistently across documents.

Comfort with unfamiliar layouts like vertical text and multi-column newspapers.

Experience with AI training data annotation, labeling, or bilingual evaluation.

Skills in transcription and localization, including MTPE, subtitling, and OCR correction.

Document production experience such as typesetting, copy-editing, or proofreading.

Knowledge of script and encoding, including Unicode normalization and Japanese input methods.

PS: Our team reviews applications daily. Please complete your AI interview and application steps to be considered for this opportunity.