Azure Document AI and Data Engineer

Role & responsibilities

Position summary

Seeking a senior Document AI and Data Engineer to implement OCR, document classification, metadata extraction, confidence scoring, structured-data, model-evaluation, and controlled retraining pipelines.

Responsibilities

• Configure OCR, layout-analysis, classification, and extraction models.
• Define schemas, labels, validation rules, and structured outputs.
• Implement file-integrity, text-layer, processability, and duplicate checks.
• Process digital documents, scanned files, images, and degraded source materials.
• Retain confidence values, page references, text spans, and source regions.
• Build normalization, transformation, and data-quality pipelines.
• Create training, validation, test, and golden datasets.
• Evaluate model accuracy, corrections, missing fields, and false predictions.
• Manage model versioning, promotion, rollback, and retraining.
• Preserve source, model, dataset, and reviewer provenance.

Required skills

• Eight or more years in data engineering, document AI, OCR, machine learning, or related data-processing disciplines
• Azure AI Document Intelligence Studio, APIs, SDKs, and custom models
• Read, Layout, classification, template, and neural extraction models
• PDF analysis, text-layer detection, image preprocessing, and degraded-scan handling
• Python, JSON, SQL, Blob Storage, Service Bus, Functions, and Azure SQL
• Document labeling, schema design, field validation, and source-region mapping
• Confidence, false-positive, false-negative, and missing-field analysis
• Golden datasets, model registries, versioning, regression testing, promotion, and rollback
• Provenance, model lineage, dataset lineage, and reviewer-correction tracking
• Production pipeline monitoring, exception handling, and performance tuning

Key deliverables

• OCR and document-processing pipeline
• Classification and metadata-extraction models
• Metadata schemas, validation rules, and structured outputs
• Golden training and evaluation datasets
• Model registry and controlled retraining framework

Preferred qualifications

• Production Azure AI Document Intelligence experience
• Strong Python, SQL, Azure data-engineering, and machine-learning pipeline experience
• Experience with large or historically diverse document collections
• Azure Data Engineer, Azure AI Engineer, or comparable certification.