Overview

• Project focused on designing an AI system that dynamically adjusts game difficulty in real time based on player behavior.
• System is genre-agnostic and does not rely on hardcoded logic; aims to work across various gameplay genres by adapting enemy behavior, puzzle complexity, and pacing.

Objectives & Key Tasks

• Implement player performance monitoring and prediction pipelines to infer player skill, engagement and frustration metrics in real time.
• Design and integrate adaptive modules to modify enemy AI, puzzle parameters, and overall game pacing without per-genre hardcoding.
• Create custom difficulty profiles and rules that the system can blend or switch between based on model outputs.

Technologies, Tools & Methods

• Primary tech stack: Unreal Engine 5 for prototype/game integration and Gameplay Ability System for in-engine adaptation.
• ML/Data stack: Python and Machine Learning models for behavior prediction and decision logic; analytics dashboard for visualizing difficulty profiles and telemetry.
• Deliverables include: prototype integration in UE5, trained models for difficulty prediction, a configurable difficulty framework, and an analytics dashboard.

Expected Outcomes & Evaluation

• Demonstrable adaptive system that changes difficulty in real time and improves player experience metrics (retention, challenge, flow).
• Documentation and examples showing how to plug the framework into different game genres and tune custom difficulty profiles.

How to Apply

• To apply, send your CV and a short motivation describing relevant ML/gameplay experience to recruitment@lanterns-studios.com .
• Use the email subject: "Application: 01 AI Universal Dynamic Difficulty Adjustment System Overview PFE" when applying.