Game AI Programming Portfolio

Ember's Oath

Game AI Programmer - Graduate Capstone Project

Team Size: 5

Tools: Unreal Engine 5, C++, Perforce

Project Length: 9+ Months (May 2025 - Present)

Project Contributions:

Programmed enemy behavior with a hybrid Behavior Tree and Utility System approach for combat gameplay. This was done in a designer-friendly, data-oriented way inspired by the Infinite Axis Utility System from Guild Wars 2.
- My system uses UCurveFloat data types as response curves, and a custom Utility Selector Composite node with custom Utility Behavior Tree Task nodes to select from after calculating consideration scores.
To learn more about how I created the Utility System's architecture, you can read about it here in this easily digestible technical document that I wrote here!
You can also view a presentation I made for my Utility System down here 👇

Capstone Utility System

Project Contributions (Continued):

Created custom EQS test cases to allow for sine scored distance and dot product test cases since Unreal Engine 5 does not natively support EQS sine scoring functions.
- This led to a more natural feeling strafing around the player.
- You can read more about how I did this here.
Created a designer-friendly Herd Manager and Enemy Spawning system which allows for rapid iteration of creating combat encounters.
- Designers connect spawn points to different waves, decide when the next wave should spawn, and add which enemies they want to spawn through a TMap where they choose an enemy archetype (the key) and the amount of that type which they want to spawn (the value).
- The herd manager keeps track of all enemies and gives permission for enemies to attack via a combat token system. Once all enemy agents are dead, it broadcasts a message to the player to reward XP/ progress the quest.
Worked on friendly, interactable NPCs that can converse with the player. Each NPC has a custom DataAsset for their specific dialogue lines.

Game AI Programmer - Rapid Prototype

Team Size: 5

Tools: Unity, C#, Git

Project Length: 1 Month (April 2025 - May 2025)

Project Contributions:

Implemented A* pathfinding using the Manhattan heuristic for enemy AI navigation in this 2D grid-based, turn-based, tactical RPG prototype.
Created melee and ranged enemy archetypes that would use a state machine to determine if it was their turn, if they should move, if they should attack, and if they should end their turn.

Game AI Programmer - Rapid Prototype

Team Size: 5

Tools: Unity, C#, Git

Project Length: 3 Weeks (March 2025)

Project Contributions:

Rapidly prototyped a 2D boss battle in which the enemy agent uses an N-Gram algorithm to gradually predict and eventually counter the player's attack patterns.

Page updated

Google Sites

Report abuse