Senior ML Engineer/Researcher

Jordan martorell s.l.; España; A distancia; EUR 50.000 - 70.000; Jornada completa

Descripción de la vacante

A leading tech company is seeking an expert in OCR and document intelligence to work remotely from Spain. The role involves researching and fine-tuning open-source models for PDF conversions and requires extensive experience in Machine Learning, specifically in OCR and vision-language models. Candidates should possess strong skills in Python and PyTorch, along with a solid understanding of document processing. This position offers an exciting opportunity to contribute to cutting-edge technology in a flexible work environment.

Formación

• 5+ years of experience in Machine Learning, with at least 2+ years focused on OCR, Document AI, or vision-language models.
• Strong hands-on expertise with Python, PyTorch, and Hugging Face Transformers (training, fine-tuning, inference).
• Practical experience deploying LLM / VLM models on high-performance inference frameworks.

Responsabilidades

• Research and fine-tune open-source OCR models for complex PDFs.
• Develop solutions for PDF-to-Markdown / PDF-to-HTML conversion.
• Build tools for data preprocessing and quality evaluation of OCR outputs.

Conocimientos

Python

vLLM

Computer Vision

PyTorch

Educación

B2 level of English

Herramientas

Hugging Face Transformers

Descripción del empleo

Location: Remote from Spain (Spanish employment contract)

We are actively experimenting with OCR and metadata extraction from the PDF documents. OCR is one of the very hot topics these days with open models actively competing for the leading places - DeepSeek OCR, LightOn OCR, etc.

We are looking for someone with the experience of running OSS models on vLLM with focus on document intelligence - computer vision that results in PDF > Markdown or PDF > HTML conversion with high precision for complex documents

Requirements

• Python
• vLLM
• Computer Vision
• PyTorch
• 5+ years of experience in Machine Learning, with at least 2+ years focused on OCR, Document AI, or vision-language models.
• Strong hands‑on expertise with Python, PyTorch, and Hugging Face Transformers (training, fine‑tuning, inference).
• Practical experience deploying LLM / VLM models on vLLM or equivalent high‑performance inference frameworks.
• Solid understanding of OCR pipelines, layout parsing, and document structure recognition (PDFs, scanned docs, tables, mixed content).
• Understanding of cloud infrastructure and GPU‑based inference pipelines.
• Research mindset with the ability to experiment, analyze, and iterate quickly.
• Strong communication and documentation skills; ability to clearly present findings and proposed improvements.
• At least B2 level of English.

Responsibilities

• Research, evaluate, and fine‑tune open‑source OCR and document intelligence models for text and layout extraction from complex PDFs.
• Develop end‑to‑end solutions for PDF‑to‑Markdown / PDF‑to‑HTML conversion with high accuracy in text structure, formatting, and layout retention.
• Build tools for data preprocessing, annotation, and quality evaluation of OCR outputs.
• Implement techniques for post‑processing, text alignment, and metadata extraction to enhance model precision.
• Collaborate with research and engineering teams to integrate OCR pipelines into production‑grade systems.
• Stay up to date with the latest developments in document AI, multimodal learning, and OCR research.