About the Position

Stage Innovation : Ingénieur Intelligence Artificielle / Computer Vision

Sèvres, IDF, France

About the Role

At the Innovation Direction, we conduct research projects across a wide range of multi-sectoral experimentation fields.

These innovative projects, developed in teams, are supervised by our experts within the ALTEN Labs (IDF, Toulouse, Grenoble, Rennes, and Sophia Antipolis), and aim to address our clients' challenges by providing them with original and disruptive technological solutions.

Within our Sèvres Lab, you will be accompanied by an Innovation Pilot (Project Manager) to help you develop your skills on one of the following projects (multiple internships available).

Projects

Project: AI for Political Debate Analysis

This project continues a platform for analyzing political debates based on AI: speaker detection, thematics, emotions, information synthesis... To promote accessibility of educational, media, and public content to deaf individuals, the goal is to add an accessibility feature by translating debates into French Sign Language (LSF) using a 3D avatar.

The objective is to develop an AI prototype capable of: Transcribing audio/video content into text, and translating it into LSF.

The result of this part will be transferred to another team within the Lab that animates a 3D avatar signing in real-time, with facial expressions and fluid gestures.

The challenge for this part is the smooth translation into LSF from text/audio, and also the integration into our AI platform for analyzing debates.

Project: Recognition of Affordances on Industrial Objects

Robotic interaction with objects relies on the ability to identify functional zones: where to grasp, push, turn, or avoid touching. This internship aims to develop an intelligent system capable of identifying the affordances of an object from an image, to determine how to manipulate it.

To this end, you will be required to:

Develop a Machine Learning model capable of segmenting affordance zones on images of industrial objects
Evaluate the robustness of the model on a real dataset
Propose a solution to take context into account to prioritize one type of manipulation over another

Project: Detection of Environmental Waste by Drones

The effective management of environmental waste has become a major challenge for the preservation of our planet. Faced with the growing pollution of waste, whether wild or from various origins, innovative solutions are necessary to preserve our ecosystems and promote sustainable development.

Traditional methods of waste detection have limitations in terms of efficiency, coverage, and cost. The tasks to be carried out are as follows:

Conduct a comprehensive review of the state of the art in environmental waste detection, focusing on methods using drones
Explore and analyze our database
Explore the application of federated learning to enable collaboration between drones and reduce the need for centralized data storage
Participate in the design and development of a drone-based environmental waste detection system, taking into account multi-objective optimization

Project: AI Application for Operator Analysis

Creation and Development of an AI:

Use of Neural Networks (transformer RNN, CNN)
Creation and adaptation of existing architectures
Use of reinforcement learning training methods
Optimization of control (frequency, data used)

An industrial application around process control by camera:

Stereo / monocular camera
Detection of operator intention and gesture
Detection and prevention of WMS (Work-Related Musculoskeletal Disorders)
Comparison of results with the manufacturing range
Use of results for the configuration of a workstation
Integration of results in a training factory

An innovation project including:

A review of the state of the art
Formalization of research with a Scientific Report
Project / multidisciplinary team management throughout the internship
An evolutionary approach (AGILE) based on Test & Learn

Requirements

You are a student in the final year of an Engineering School looking for a final year internship and have specialized in Signal/Image Processing, Computer Science, or Artificial Intelligence.

You have knowledge in Computer Vision, computer development, or Machine Learning, algorithms, and Artificial Intelligence, and are proficient in several associated tools such as PyTorch / Tensorflow, OpenCV, object recognition algorithms (Yolo, SSD), as well as the Python or C++ languages, which you have applied during your training.

Responsive, rigorous, autonomous, and with a sense of service, you wish to evolve in a challenging environment.

How to Join One of Our Labs?

Share your CV so that we can discover your background and what drives you.
Exchange with a recruitment officer during an initial discussion to better understand your expectations.
Meet an innovation pilot to talk about concrete projects, ideas... and see if the magic happens.

Joining our ALTEN Labs means immersing yourself in our culture of innovation. You will have the opportunity to develop your skills on concrete subjects, collaborating within a dynamic project team.

Our teams of experts will support you to become an actor in your project, in a multicultural and multidisciplinary environment. You will thus benefit from an enriching experience, open to all sectors of engineering, in France and internationally.

Internship Details

Duration: 6 months (A job offer may be considered at the end of the internship)
Location: Sèvres Lab
Stipend: 1300 euros gross/month + Meal vouchers + Public transport reimbursement

Do you see yourself in this description? Then don't wait any longer! Join a team that places innovation at the heart of its actions and where every talent can leave its mark.

Stage Innovation : Ingénieur Intelligence Artificielle / Computer Vision

Job description