Boss Fight - Project SmartEnemyAi - Unreal Engine 5.6 & Blueprints

Smart Enemy AI System (Combat-Focused, Modular Design) Developed a modular and scalable enemy AI system designed to support diverse combat behaviors across multiple enemy types, including melee, ranged, magic, and boss units. The system features a state-driven architecture with five core states: Passive, Investigation (with active seeking), Attacking, Frozen, and Death. Enemies dynamically transition between these states based on environmental stimuli, player interaction, and internal decision logic. Each enemy type is equipped with distinct combat capabilities: Melee Enemies: Utilize a variety of attacks such as ground smash, sword slashes, jump attacks, and combo chains. Ranged Enemies: Engage from distance using projectile-based attacks. Magic Enemies: Implement advanced behaviors including teleportation (with VFX), area-of-effect attacks, and targeted spell casting. Boss Enemies: Combine all enemy archetypes with extended abilities such as multi-phase combos, weapon throws, and alternate invincible attack states. A token-based attack system is implemented to regulate combat flow, ensuring only a controlled number of enemies can attack the player simultaneously, improving gameplay balance and realism. The AI also supports team-based logic, preventing friendly fire and enabling coordinated group behavior. Combat interactions are enhanced with shared systems such as: Intelligent attack selection, where enemies choose actions based on context (distance, player state, etc.) Blocking and parrying mechanics Damage handling via reusable components The overall architecture is built using a general AI interface and modular components (e.g., Combat, Damage), making the system extensible and easy to integrate into different gameplay scenarios.