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