Python Backend Developer

Experteer Overview

You’ll lead comparative evaluations of OCR and AI models and implement end-to-end workflows used by a global team. This role blends strategic decisions with hands-on development in a fast-paced, startup-like AI environment.

• Conduct in-depth comparative analyses of commercial and open-source OCR/AI solutions (ABBYY, Adobe, AWS Textract vs. Mistral OCR, Docling, Nougat, LlamaParse)
• Define and measure format fidelity metrics for headers, footers, tables, and styles
• Develop a Python-based workflow integrating OCR engines with document-generation libraries (python-docx, Pandoc)
• Experiment with Vision-Language Models and LayoutLM-style architectures to improve structural recognition
• Address edge cases (rotated text, low-res scans, complex mathematical notation) to optimize the pipeline
• Guide the document conversion research agenda and balance build vs. buy decisions (off-the-shelf API vs. custom PyTorch/TensorFlow pipelines)

• Expert-level Python with OpenCV, PyMuPDF, and python-docx
• Deep familiarity with Tesseract, PaddleOCR, LayoutLMv3, Donut, Nougat
• Strong understanding of XML-based formats (OOXML) and pixel-perfect formatting
• Experience integrating LLMs (GPT or Claude) for layout correction and semantic cleanup
• Architectural judgment for when to leverage off-the-shelf APIs versus custom AI pipelines (PyTorch/TensorFlow)

•