OCR/IDP Data Labelling & Validation Specialist - Contract -

Important Note

This is a project-based contract role with an initial 6-month duration. While contract extensions may be offered based on performance and business needs, this role does not convert to full-time employment unless explicitly stated.

Position Overview

We are seeking detail-oriented Data Labeling & Validation Specialists to support ABBYY's OCR and Intelligent Document Processing (IDP) systems.

This role combines hands-on document annotation with structured validation of automated labeling outputs. You will play a key role in the human-in-the-loop pipeline, ensuring machine learning models are trained on high-quality, accurate ground truth data.

Success in this role requires prior hands-on annotation experience and the ability to evaluate whether automated outputs meet quality expectations, identify error patterns, and provide structured feedback to improve model performance.

Key Responsibilities

Document Annotation

• Annotate semi-structured and unstructured documents across diverse formats and domains

• Perform labeling across key IDP elements, including:

• Text recognition (including handwriting)

• Document classification

• Field extraction (PII, dates, amounts, signatures, etc.)

• Table detection and structure

• Label document layout elements such as zones, reading order, and hierarchy

• Verify OCR output accuracy and correct recognition errors

• Handle complex or ambiguous document formats beyond automated capabilities

• Maintain high levels of accuracy and consistency across all annotation tasks

Auto-Label Validation & Error Analysis

• Review sampled subsets of auto-labeled outputs and validate against ground truth

• Identify, categorize, and document errors-including distinguishing:

• Isolated issues

• Systematic failure patterns across document types

• Provide structured, actionable feedback to ML engineering teams

• Assess confidence scores and flag outputs below quality thresholds

• Track validation metrics over time and identify quality trends

Quality Assurance & Feedback

• Review annotations completed by other team members to ensure consistency

• Identify and document edge cases (e.g., unusual layouts, ambiguous fields)

• Participate in calibration sessions to align on annotation standards

• Provide feedback to improve annotation guidelines and workflows

• Adhere strictly to data privacy and confidentiality standards

Qualifications

Education & Experience

• High school diploma or equivalent; Associate's or Bachelor's degree preferred

• 1+ year of hands-on experience in document annotation or data labeling (direct annotation required)

• Proven ability to maintain high accuracy in repetitive, detail-oriented tasks

• Experience working with and following annotation guidelines

Technical Skills

• Familiarity with annotation tools and labeling platforms

• Understanding of document structure and layout types

• Basic knowledge of data privacy and security practices

• Reliable computer and high-speed internet connection

• Strong English reading comprehension and written communication skills

Analytical Skills

• Ability to distinguish between isolated errors and systematic issues

• Strong pattern recognition across large datasets

• Critical thinking to evaluate ambiguous cases and escalate appropriately

• High attention to detail when reviewing auto-generated outputs

Preferred

• 1-2 years of experience in OCR, IDP, or document labeling workflows

• Experience with auto-labeling systems or AI-assisted annotation tools

• Background reviewing or auditing machine-generated outputs

• Familiarity with inter-annotator agreement and data quality metrics

• Domain expertise in document-heavy industries (e.g., finance, legal, healthcare)

• Proficiency in languages beyond English

• Experience with spreadsheets, data tracking, or reporting tools

Compensation & Benefits

• Competitive hourly rate (based on location and experience)

• Flexible schedule within project deadlines

• Remote work environment

What You'll Gain

• Hands-on experience with real-world AI/ML data pipelines

• Direct collaboration with machine learning engineers

• Exposure to auto-labeling systems and document AI technologies

• Development of skills in data quality, validation, and error analysis

• Experience valuable for future roles in ML data operations, QA, or annotation engineering

Training & Support

• Structured onboarding (1-2 weeks) covering tools, workflows, and guidelines

• Ongoing support from project managers and technical teams

• Access to detailed documentation and best practices

• Regular performance feedback with metrics and improvement insights

Project Details

• Duration: 6-month contract (renewal based on performance and project needs)

• Workload: Typically 20-40 hours per week depending on project phase

• Team Structure: Distributed team with established communication channels

• Performance Metrics:

• Annotation accuracy

• Validation throughput

• Quality of error documentation

• Adherence to guidelines

Application Requirements

Please submit:

• Resume highlighting relevant annotation, data labeling, or QA experience

• Cover letter describing your approach to identifying errors in automated outputs

• Work samples (if available) demonstrating document labeling or review accuracy