Senior OCR Evaluation Engineer — Remote

Himpunan Profesi Hubungan Industrial (HPHI); United Arab Emirates; A distancia; AED 260.000 - 380.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

100% remote

Competitive compensation

Professional development

Collaborative, international team

Descripción de la vacante

Himpunan Profesi Hubungan Industrial (HPHI) is seeking a Senior Optical Character Recognition (OCR) Evaluation Specialist for a fully remote role. You will design evaluation methods and benchmarks to assess OCR accuracy, reliability, and performance across diverse documents and languages.

The ideal candidate has deep experience in OCR evaluation, data science, and document analysis, with strong Python skills and familiarity with major OCR tools.

Formación

• Bachelor's or master's degree in Computer Science, AI, Data Science, computational linguistics, IT, or related field.
• Significant experience in OCR evaluation, document processing, computer vision, or ML QA.
• Strong understanding of OCR tech, text recognition, and document image processing.
• Experience evaluating OCR engines and benchmarking recognition performance.
• Knowledge of CER, WER, edit distance, and statistical eval methods.
• Proficiency in Python and image processing/data analysis libraries.
• Familiarity with OCR tools like Tesseract, PaddleOCR, EasyOCR, or Azure AI Document Intelligence.
• Experience with OpenCV and Pillow.
• Understanding of document layout analysis, text segmentation, and table recognition.
• Experience preparing ground-truth datasets and annotation QC.
• Knowledge of multilingual text recognition is advantageous.
• Familiarity with ML model evaluation, regression testing, and performance monitoring.
• Strong analytical reasoning, troubleshooting, and attention to detail.
• Excellent technical documentation and communication skills.
• Ability to work independently and remotely.

Responsabilidades

• Design and implement evaluation frameworks for OCR systems and document recognition tech.
• Assess OCR accuracy across formats, languages, fonts, layouts, and image conditions.
• Develop benchmark datasets containing scanned documents, images, and ground-truth text.
• Evaluate CER, WER, precision, recall, and other relevant performance metrics.
• Compare OCR engines, configurations, and models against quality benchmarks.
• Analyze recognition errors including substitutions, omissions, insertions, formatting, and segmentation.
• Evaluate performance on low-resolution, skewed, noisy, handwritten, and complex layouts.
• Assess preprocessing techniques like enhancement, binarization, deskewing, and noise removal.
• Validate text extraction accuracy for tables, forms, multi-column layouts.
• Conduct linguistic and contextual reviews to catch errors automated metrics miss.
• Develop automated testing scripts and QA procedures for OCR evaluation.
• Investigate performance variations across languages, scripts, document categories, and sources.
• Collaborate with ML engineers, document intelligence specialists, data scientists, and software developers.
• Recommend improvements to OCR models, preprocessing pipelines, and post-processing methods.
• Establish quality thresholds, acceptance criteria, and regression testing standards.
• Maintain evaluation datasets, test docs, error taxonomies, and performance reports.
• Monitor OCR model performance after updates and releases.
• Present evaluation findings and recommendations to stakeholders.

Conocimientos

Python

OCR evaluation

OpenCV

Pillow

Educación

Bachelor's degree

Master's degree

Herramientas

Tesseract

PaddleOCR

EasyOCR

Azure Document AI

Descripción del empleo

Himpunan Profesi Hubungan Industrial (HPHI) is seeking a Senior Optical Character Recognition (OCR) Evaluation Specialist for a fully remote role. You will design evaluation methods and benchmarks to assess OCR accuracy, reliability, and performance across diverse documents and languages.

The ideal candidate has deep experience in OCR evaluation, data science, and document analysis, with strong Python skills and familiarity with major OCR tools.