Research Intern, Docling Team

IBM Research Zurich seeks a Research Intern for the Docling Team to advance document AI through model development, algorithm design, and research-to-production collaboration.

This internship places candidates within the Docling Team at IBM Research Zurich, focusing on the full stack of document AI research and development. The position provides exposure to research-to-product pipelines and the broader enterprise AI ecosystem, allowing interns to contribute to open-source projects and translate academic findings into production-ready solutions.

Responsibilities

• Participate in the training, evaluation, optimization, and large-scale deployment of document AI models.
• Experiment with vision-language models for tasks such as document layout analysis, OCR, information extraction, and multimodal understanding.
• Design and implement novel algorithms that combine visual, textual, and structural signals to improve document understanding.
• Work with large-scale datasets and training pipelines using modern deep learning frameworks.
• Collaborate with researchers and engineers to translate research into production and contribute to open-source projects.
• Analyze experimental results, run ablations, and iterate methods to advance state of the art in document intelligence.

Requirements

• Currently enrolled in or recently completed a PhD or Master's with research experience in Computer Science, Machine Learning, or a related field.
• Strong foundation in machine learning, deep learning, computer vision, natural language processing, or multimodal learning.
• Proficiency in Python and modern deep learning frameworks (e.g., PyTorch) and libraries such as Hugging Face Transformers / Datasets.
• Excellent communication skills (written and verbal), ability to present and write clearly.

Nice to have

• Publications in relevant domains (CVPR, ICCV, ECCV, NeurIPS, ICLR).
• Experience with document-centric tasks: OCR, layout modeling, table understanding, DocQA.
• Familiarity with large-scale training, distributed computing, or model optimization.
• Experience working with pre-trained models, fine-tuning, multimodal models or architectures.

What the company offers

• Outstanding performance may lead to future employment opportunities within the Docling team.
• Excellent flexible working arrangements.

About the company

IBM is committed to diversity at the workplace. The environment is open and multicultural.

• Commitment to diversity at the workplace.
• Open and multicultural environment.