Intership Agentic Multimodal AI for Understanding and Reasoning over Complex Technical Documents-H/F

CEA; Saclay; On-site; EUR 10,000 - 12,000; Part time

Job summary

Le LIST du CEA à Saclay propose un stage en IA multimodale agentique pour la compréhension et le raisonnement sur des documents techniques complexes. Le stagiaire travaillera sur l’évaluation de modèles et l’orchestration d’outils pour décomposer des requêtes et fusionner des informations issues de texte, tableaux et figures sur plusieurs pages.

Ce stage offre l’accès à un environnement de recherche de haut niveau et à des ressources avancées, avec la possibilité de publier des résultats et

Qualifications

• Étudiants en 4e/5e année (M1, M2 ou année sabbatique).
• Intérêt fort pour l’IA multimodale générative.
• Connaissances des LLM, VLM, Transformers, RAG et systèmes agentiques.
• Compétences en Python avec un framework DL (TensorFlow ou PyTorch).

Responsibilities

• Examiner des benchmarks existants et développer un benchmark si nécessaire.
• Évaluer des VLM récents pour la compréhension de documents techniques.
• Concevoir une architecture agentique pour décomposer des requêtes et planifier un raisonnement multi-étapes.
• Récupérer dynamiquement des pages et sélectionner outils ou modèles pertinents.
• Combiner et raisonner sur des informations provenant de texte, tableaux et figures sur plusieurs pages.
• Évaluer l’approche proposée par rapport à des baselines VLM et multimodales.

Skills

Multimodal IA

LLMs

Vision-Language Models

Transformers

RAG

Agentic systems

Computer vision

Deep learning

Python

Education

4e/5e année (M1/M2 ou année sabbatique)

Job description

Intership Agentic Multimodal AI for Understanding and Reasoning over Complex Technical Documents-H/F

Domaine

Contrat

Stage

Intership Agentic Multimodal AI for Understanding and Reasoning over Complex Technical Documents-H/F

The internship aims to develop multimodal agentic AI for understanding and reasoning over complex technical documents containing text, tables, figures, diagrams, and equations distributed across multiple pages. The work will focus on evaluating recent Vision-Language Models (VLMs), designing an agentic architecture capable of decomposing complex queries, planning multi-step reasoning, and dynamically selecting relevant pages, tools, and models, and combining information across multiple modalities. A dedicated benchmark may also be developed, and the proposed approach will be evaluated against direct VLM and multimodal RAG baselines.

As an intern at the CEA, you will have the opportunity to work within a world-renowned research environment. Our teams consist of passionate and dedicated experts, providing an ideal setting for learning and collaboration. You will have access to state-of-the-art equipment and top-tier research resources to carry out your assignments. The work conducted may potentially lead to a scientific publication.

Context :

Complex technical documents combine heterogeneous information such as text, tables, diagrams, figures and equations, often distributed across multiple pages. They include technical manuals, maintenance procedures, user guides, datasheets, product specifications, engineering documents, and technical or scientific reports.

Understanding such documents raises several challenges: retrieving relevant information across long documents, jointly interpreting heterogeneous modalities, linking information distributed over multiple pages, and performing multi-step reasoning.

This internship aims to develop multimodal agentic AI capable of planning and orchestrating specialized models and tools for complex document understanding and reasoning.

What do we expect from you?

To address these challenges, the internship will investigate the following research directions:

• Review existing benchmarks and, if necessary, develop a benchmark covering a fine-grained taxonomy of technical documents and multimodal reasoning tasks.
• Evaluate recent Vision-Language Models (VLMs) for technical document understanding.
• Design an agentic architecture able to decompose complex queries and plan multi-step reasoning.
• Dynamically retrieve relevant pages and select appropriate tools or models.
• Combine and reason over information from text, tables, figures, diagrams and equations across multiple pages.
• Evaluate the proposed approach against direct VLM and multimodal RAG baselines.

#Cea-List

• Students in their 4th or 5th year of studies (M1, M2 or gap year)
• Strong interest for Multimodal Generative AI
• Knowledge of LLMs, Vision-Language Models (VLMs), Transformers, RAG and agentic systems
• Computer vision skills
• Machine learning skills (deep learning, perception models, generative AI…)
• Python proficiency in a deep learning framework (especially TensorFlow or PyTorch)

Site

Saclay

Langues

Oui

Référence

2026-41999

Based in Saclay (Essonne), the LIST is one of the two institutes of CEA Tech, the Technological Research Division of the CEA. Dedicated to intelligent digital systems, its mission is to carry out technological developments of excellence on behalf of industrial partners, in order to create value.
Within the LIST, the Laboratory of Vision and Learning for Scene Analysis (LVA) conducts its research in the field of computer vision and artificial intelligence for the perception of intelligent and autonomous systems. The laboratory's research themes include visual recognition, behavior and activity analysis, large-scale automatic annotation, and perception and decision models.