Trainings

State of the Art in Technology and Science: World Models (2 days)

World Models emerged in 2018 as a more robust approach to Deep Reinforcement Learning and were subsequently popularized by figures such as Yann Lecun. Spanning the learning of a system’s latent dynamics and their application to generative models or agent control (in the context of DRL), this field heralds the next revolution in Artificial Intelligence, with numerous applications already in use. This course starts with the fundamentals (VAE, model-based reinforcement) to introduce foundational research and various modeling approaches, and goes on to present existing applications, future opportunities, and the limitations of these tools as currently understood.

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Fundamentals: Deep Learning, Mathematics

Durée

2 days
State of the Art in Technology and Science: Cybersecurity and AI (2 days)

Artificial intelligence has ushered in a new paradigm in system security, bringing with it a whole range of new risks: adversarial attacks, poisoned datasets, model backdoors, model inversion, prompt injections, and jailbreaks. Based on recent academic research, this course presents these types of attacks in detail, before moving on to discuss AI-based application approaches. Protection and mitigation methods are proposed, along with their risk-benefit ratios, to provide a comprehensive overview.

Public

Ingénieurs et programmeurs, data-engineers

Pré-requis

Fundamentals: Cybersecurity, AI Tools

Durée

2 days
State of the Art in Technology and Science: Large Language Models & Agentic (2 days)

Large Language Models have largely established themselves as the central architecture of AI since the emergence of the Transformer. Numerous studies have sought to improve these tools: model architecture, advancements in training methods (and the introduction of reinforcement learning), the shift toward agent-based systems, and more. This course aims to provide the scientific and technical foundations needed to gain a solid understanding of current LLM and agent-based architectures and their training methods. It also delves into topics such as model optimization and safety, and then explores major upcoming developments (Diffusion, Mamba).

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Fundamentals: Deep Learning, Data Science

Durée

2 days
AI Strategy and Governance (2 days)

From managing an AI project to implementing these tools within an organization, it is essential to fully understand the opportunities and limitations of these tools in order to manage risks and ensure the success of projects beyond a simple proof of concept. Drawing on numerous publications dedicated to the subject (LVMH Application, Ethical AI, AI Act, economic mastery, etc.) from reputable sources (CRJ CE, Stanford, ISO 42001 standard), this training course aims to provide the tools needed to understand the challenges, avoid common pitfalls, and turn projects into well-managed investments.

Public

Tout public au contact de l'IT

Pré-requis

Management/Leadership Experience

Durée

2 days
State of the Art in Technology and Science: Diffusion Models (2 days)

Diffusion models have established themselves as an essential framework in many fields (image, audio, reinforcement learning, language), particularly due to their high level of generalizability. This course provides an understanding of these models, their specific characteristics, and their applications. It covers the fundamental research and models published since 2019 and integrates them into a simplified, comprehensive overview. It then presents the various applications, concluding with the scientific community’s current understanding of what explains the high effectiveness of these models.

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Fundamentals: Deep Learning, Mathematics

Durée

2 days
Project Management and Control: LLMs and Agent-Based Systems (1 days)

LLM and agent-based tools are indispensable, but they remain difficult to master due to their novelty and the field’s constant evolution. The challenge is to provide a well-reasoned and substantiated overview—free of fantasy—to offer the keys to managing and steering a project based on these technologies. Managing “hallucinations,” controlling costs, choosing a model, testing an AI tool, and cybersecurity challenges are among the topics covered during this event and illustrated by research findings and real-world experience.

Public

Tout public au contact de l'IT

Pré-requis

Project Management/Leadership Experience

Durée

1 days
State of the Art in Technology and Science: AI and Computer Vision (2 days)

While Large Language Models are currently the focus of attention, artificial intelligence has also been, since 2012, a promising path (albeit not without its pitfalls) for applications in images, 3D models, and video. This course aims to provide a comprehensive overview of the technical and scientific foundations of the various approaches and implementations in image-based deep learning as of 2026. Covering fundamental architectures (CNN, Vision Transformer, NeRF, Diffusion), the course also presents the main fields of application, highlighting both the advantages and limitations of these approaches.

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Fundamentals: Deep Learning, Data Science, Image Processing

Durée

2 days
Deep learning and sound processing training (3 days)

With the success of voice assistants such as Alexa/Siri/Google, deep learning has brought sound and speech processing into the mainstream. Topics include speech recognition, denoising, classification, audio tagging and audio separation (speech & music)…

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Python programming basics

Durée

3 days
Deep learning and image processing training (2 days)

Image processing is one of the fields that has benefited most dramatically from advances in Deep Learning. Topics range from classification and segmentation to image transformation, including the generation of text-oriented analyses.

Public

data-scientists, data-engineers

Pré-requis

Python programming basics, mathematics bac+3

Durée

2 days
Deep learning and language processing training (3 days)

Deep Learning has recently revolutionized natural language processing: translation, feature identification, dialogue systems, interpretation, etc.

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Python programming basics, mathematics bac+3

Durée

3 days
Deep reinforcement Learning training (2 days)

Reinforcement learning aims to teach an agent how to optimize its actions so as to maximize its gains. This classic field has recently been revolutionized by Deep Learning (Q function, policy, etc.) and now makes it possible to solve tasks previously considered beyond the machine’s reach: process optimization, go/video games, robotics, etc.

Public

data-scientists, data-engineers

Pré-requis

Python programming basics, mathematics bac+3

Durée

2 days
Pytorch training (3 days)

After revolutionizing many scientific fields, artificial intelligence is now taking root in industry. In practice, it is supported by Python frameworks such as PyTorch, now a leader in deep learning.

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Python programming basics

Durée

3 days
Python training for Deep Learning (3 days)

Data mining has become an integral part of R&D activities. However, understanding the issues and the state of the art in the field requires a solid mathematical background.

Public

Ingénieurs et programmeurs, data-scientists

Pré-requis

Python programming basics

Durée

3 days
Introduction to AI: Tools and challenges (2 days)

Artificial intelligence has revolutionized many fields of science, and is now beginning to revolutionize a large number of economic sectors (industry, medicine, communications, etc.). Nevertheless, its presentation in the mainstream media is often nothing more than fantasy. The aim of this training course is to present these approaches and their contribution to solving problems considered to be “intelligent”.

Public

Tout public au contact de l'IT

Pré-requis

IT project management

Durée

2 days
Mathematics for data scientists and artificial intelligence (3 days)

Data mining has become an integral part of the business of R&D. However, understanding the issues and the state of the art in the field requires a solid mathematical background.

Public

data-scientists, data-engineers

Pré-requis

Notions of statistics, probability, mathematical background

Durée

3 days