Thibaut Platet

• Developed in C++ (Unreal Engine 5), within the LSI’s proprietary framework, the VR visual guides of an industrial assembly training application built on the XDE physics engine: “ghost” parts replay the degrees of freedom of each mechanical joint to guide the operator.
• VR interaction: teleportation, grabbing physics-driven objects via XDE manipulators, Blueprint interactions and abstractions.
• Design, development and debugging in an Agile workflow.
• Wrote the module’s technical documentation, shared across the laboratory.

• C++
• Unreal Engine 5
• VR
• Physics simulation (XDE)
• Blueprints
• Agile

Environmental and anthropological simulation in AR, driven by Deep Reinforcement Learning (CNN).

Problem

Show the consequences of a planning choice (pollution, floods) before anything is built: a sustainable city simulation designed for urban planning studies.

Role

Project lead, team of 3.

Decisions

An AI trained with Deep Reinforcement Learning (CNN), learning in stages; a Python simulation engine linked over UDP to a Unity app in augmented reality.

Result

Working prototype, tried by visitors at the ENSIM open day. Over 1,000 games, the AI scores on average 4.75 times higher than a random strategy.

• AR
• Deep Reinforcement Learning
• Unity
• Python

Survival mini-game in Unreal Engine 5, built entirely in Blueprints.

Problem

Escape from an industrial complex overrun by zombies, with limited time and limited health.

Role

Solo project.

Decisions

Waves that get fiercer and fiercer, doors that only open after certain waves, two pick-ups (health and time) and a hidden code to type on a keypad for the final door.

Result

A playable game, shown on video: nine waves to hold off, five doors to get through and a final puzzle.

• Unreal Engine 5
• Blueprints

Serious game for learning astronomy in VR.

Problem

Give a young audience some scientific culture about the solar system, using the immersion of virtual reality to convey the vastness of space.

Role

Project lead, team of 4.

Decisions

A fixed route from planet to planet rather than free exploration, which was too heavy to build; three surfaces to land on (Moon, Mars, Earth); a text and a voice-over for each planet.

Result

A complete game, from the menu to the final scene, tested on Meta Quest 3. A first version was tried by many visitors at the ENSIM open day (25 January 2025).

• VR
• Unity
• C#
• Blender

VR minigolf on an island of ancient temples, built as a pair.

Problem

Build a virtual reality minigolf game in a peaceful world: an island surrounded by the sea, where a mascot welcomes players and shows them the moves.

Role

Built as a pair. My part: the island, the menus and their interfaces, the C# scripts (score, menus, ball physics) and the mascot’s animations.

Decisions

Merge our two initial ideas into a single world; have an animated mascot demonstrate the swing; have the ball struck with a real movement, club in hand.

Result

A game in two scenes: a start menu, then the island and its three courses, all playable to the end: you pick up a club, strike the ball and track your number of strokes.

• VR
• Unity
• C#
• Blender

• Built territorial data study grids in QGIS and ArcGIS.
• Data sorting with Python and Power BI.

• QGIS
• ArcGIS
• Python
• Power BI

Carbon footprint assessment of ENSIM’s emissions, following ADEME criteria.

Problem

Measure the school’s greenhouse gas emissions over 2023, to know where to act and to raise awareness among students and staff.

Role

Project lead, team of 3.

Decisions

The GHG Protocol method in three scopes, recommended by ADEME; a survey of students and staff, completed by interviews with the school’s departments; an uncertainty estimated for every item.

Result

A full report: emission items ranked, a comparison with other schools and ways to reduce them, together with awareness posters.

• GHG Protocol
• Excel
• Online surveys (Framasoft)
• ADEME data