Development Engineer — Human-AI Real-time Interactive Systems

IMT Atlantique, a leading generalist engineering school within IMT-Institut Mines-Télécom, France's premier group of engineering schools, aims to support transitions, train responsible engineers, and leverage scientific and technical excellence for teaching, research, and innovation.

Context of the position:

The national PEPR eNSEMBLE program (France 2030) and its targeted MATCHING program (https://www.pepr-ensemble.fr/matching-collaboration-avec-des-systemes-intelligents/) focus on the study and transformation of digital collaboration environments, aiming for new socio-technical models that enable effective, inclusive, and interoperable collaboration between individuals and intelligent systems. This position is part of two targeted projects within the MATCHING program:

• DÉCAFEIN': A multi-user virtual collaborative environment (virtual reality) where participants interact with socially interactive virtual agents to solve an investigation, with multimodal analysis of behavior (verbal, non-verbal, physiological) for therapeutic, educational, or recreational purposes.

• CollabJam: A playful demonstrator of collective musical game with a generative AI as a musical partner, also designed as a scientific experimentation tool. Through various scenarios, it will allow studying different forms of human-AI collaboration, varying the AI partner's behavior (support, offering variations, or deliberate disruptions), as well as non-verbal communication methods allowing participants to share and negotiate their aesthetic intentions.

Both projects share the same scientific and technical challenge: designing a system where an agent (virtual or musical) perceives multimodal signals produced by a group in real-time and dynamically adapts its own level of autonomy or cooperation based on collective activity. The recruited engineer will contribute to the technical development of both demonstrators, prioritizing the sharing of generic modules and respecting the specificities of each.

1- MISSIONS

1. Common Foundation for both projects

• Design and develop real-time interaction loops between a group of users and an adaptive agent (virtual or generative).

• Implement the acquisition, synchronization, and time-stamping of heterogeneous multimodal streams depending on the project (audio, gestural, and physiological for DÉCAFEIN'; MIDI streams for CollabJam).

• Implement a logic for modeling the agent's 'agency level' or cooperation, driven by the group's interaction history.

• Design mechanisms for presenting the agent's state to users (feedback, visualization, sensory feedback).

• Contribute to the production of multimodal corpora usable for scientific evaluation of collaboration.

• Document developments and participate in exchanges with research teams from PC1 and PC3 of the eNSEMBLE program.

2. Specific to DÉCAFEIN'

• The virtual agents in DECAFEIN are embodied.

• Develop the multi-user virtual reality environment using Unreal Engine 5 (integration of Metahumans for virtual agents).

• Implement the distributed hardware architecture for 5 workstations (4 participants + 1 expert) and network synchronization between instances.

• Integrate capture devices (eye-tracking, heart rate, electrodermal activity, audio) and their time-stamping with simulation events.

• Implement the logic conditioning the revelation of clues by virtual agents based on the quality of multimodal behavior detected in participants.

3. Specific to CollabJam

• The virtual agent(s) in CollabJam are not embodied a priori.

• Design and develop the real-time architecture for capturing, analyzing, and exploiting the musical interactions of a group of participants (initially based on MIDI data with the possibility to explore other audio/gestural forms later).

• Develop the adaptation mechanisms for the AI musical partner to evolve its behavior between imitation, accompaniment, and disruption.

• Design and implement interaction scenarios to experiment with different forms of collaboration and negotiation of agency between users and AI, for scientific data collection.

• Ensure the fluidity of the musical experience by working on synchronization, system responsiveness, and latency management.

• Explore with project researchers sound, visual, or gestural modalities to make the AI's state and intentions perceptible, and to support non-verbal communication between participants.

Required Skills:

✔️ Real-time development: Solid experience in C++ and/or C# programming; design of interactive systems with latency constraints.

✔️ 3D / VR (desired): Unreal Engine 5, ideally with Metahumans; Unity / Godot; multi-user networking in an immersive environment.

✔️ MIDI / music (desired): Real-time MIDI stream processing (e.g., JUCE, Max/MSP, RtMidi); notions of AI-based music generation. The CollabJam project is limited to MIDI streams, without audio signal processing.

✔️ Multimodal signal processing: Acquisition and synchronization of heterogeneous streams (physiological, gestural, audio); notions of signal processing.

✔️ AI / adaptive models: Notions of adaptive agents, generative models, or online learning are appreciated.

✔️ Interpersonal skills: Autonomy, ability to work with multidisciplinary research teams (cognitive science, HCI, psychology), documentary rigor.

Profile:

🎓 Engineering degree or Master 2 in computer science, AI, computer graphics, or audio signal processing.

💼 Prior experience in at least one of the two areas (VR/Unreal Engine or MIDI streams) is an asset; the other can be acquired on the job.

💡 Interest in human-AI collaboration and social interaction issues.

Summary: Mutualization vs. Specificities

Both projects share a common foundation of skills in real-time interactive systems engineering and adaptive agent modeling, but rely on distinct technical stacks (3D/VR game engine for DÉCAFEIN', real-time MIDI pipeline for CollabJam — acoustic/audio signal processing is not within the scope of this position). Depending on the profile of the selected candidate, a work time distribution between the two projects will be defined with the respective scientific leads.

WHY JOIN US:

👉 Exceptional working environment

👉 Stimulating innovation ecosystem

👉 Attractive location: Brest Campus (Plouzané), accessible by public transport, by the bay

PERKS:

🥗 On-site collective catering

🏃‍♀️ Leisure/sports offerings

🚌 Public transport reimbursement

🚴‍♂️ Sustainable mobility package (for carpooling or cycling commutes)

👨‍👩‍👧‍👦 Family benefits

💶 Wide range of social benefits

🖥️ Partial remote work possible

🌴 Numerous holidays

REMUNERATION INDICATIONS:

💰 Indicative remuneration: from €33,648 gross per year, depending on profile and experience (reduced social charges in the public sector).

FOR ANY INFORMATION:

• On the job content: Gilles COPPIN - Professor IMT Atlantique / Lab-STICC: On administrative/HR aspects: Mélissandre MORVAN – Recruitment Assistant

Application deadline: 23/09/2026

Start of contract: as soon as possible

Interviews: on a rolling basis

Legal Mentions:

• Nature and duration of the contract: 12-month fixed-term contract under public law or detachment on contract for civil servants, renewable.

• Job location: IMT Atlantique Brest campus - 655 Av. du Technopôle, 29280 Plouzané

• For internal use:

  • Category – Job title: II R

  • Category and job title of persons eligible to apply:

    • II: D, E, F, G, H, I, R

    • III: J1, J2, K, L, M

• The positions offered for recruitment are open to all, with, upon request, accommodations for candidates with disabilities.

• Job open to civil servants and/or contract employees.

• All applications may be subject to an administrative investigation.

Find Jobs in France on Arbeitnow