STEM OCR Specialists Part-time NEW

Mathematics & Stem Textbook QA Specialist (Ocr + Latex)

Mercor is partnering with a leading frontier AI research lab on an initiative to improve math and STEM reasoning. We're building high-quality training and evaluation data from public-domain textbooks, and we're looking for detail-oriented STEM experts to validate and repair extracted problem–solution content.

This role is ideal for people who are strong at proof-based math reading, precise transcription, and LaTeX formatting, and who enjoy meticulous quality work.

Key Responsibilities

• Validate math/STEM question–answer pairs extracted from textbook PDFs, identifying errors introduced by OCR or extraction.

• Compare extracted text to the source PDF and correct transcription issues (symbols, subscripts/superscripts, missing terms, etc.).

• Fix and standardize LaTeX formatting so all expressions render correctly and match the source.

• Extract and label relevant context (problem statements, surrounding definitions, hints, solution steps, short answers) and record correct PDF viewer page numbers.

• Revise raw textbook content into clean QA format when the original is demonstrative (e.g., “Show that…”) or references missing equations/figures.

• Assess solution quality (correctness, completeness, logical consistency) and flag gaps, incorrect steps, or mismatches with the stated answer.

• Maintain consistently high standards for precision, structure, and correctness.

Required Qualifications

• Strong proficiency with LaTeX math formatting (required).

• Background in Mathematics, Physics, Engineering, or related STEM (BA/BS/Masters/PhD or equivalent rigor).

• Comfort reading and verifying multi-step mathematical reasoning (including proofs and textbook solutions).

• Excellent attention to detail and ability to follow a structured workflow exactly.

• Strong written communication for clear, concise technical notes.

Preferred Qualifications

• Proof-based coursework (e.g., real analysis, abstract algebra, proof-based linear algebra).

• Experience as a TA/grader, tutor, contest solution writer, or math editor.

• Familiarity with common OCR failure modes in STEM text (misread symbols, broken equations, malformed fractions, etc.).

• Ability to handle advanced undergraduate / occasional graduate-level material.

What You'll Be Evaluating

• Exercises and examples across math and STEM

• Proof-oriented problems

• Occasionally multi-part questions, with dependencies across subparts

• Problems that require careful handling of figures/diagrams and referenced equations

More About the Opportunity

• Potential commitment: 15 hours/week minimum, up to 40 hours/week

• Approximate project length: 2-3 weeks

• Work is structured and quality-focused; speed matters less than correctness

We consider all qualified applicants without regard to legally protected characteristics and provide reasonable accommodations upon request.