Gameplay & AI
Programmer.

As technical lead for Miasma Ashlung, a 9-week extraction horror game, I spearheaded technical development by setting up the GitHub repository, porting the prototype to C++, and leading the pivot from procedural generation to manual level design for stability. I developed core systems like Miasma mechanics, enemy AI using Unreal Engine’s Perception System, and physics-based potion throwing. I enhanced immersion with an overhauled Miasma material, customizable VFX, and audio cues with AI-reactive 'Make Noise' functionality. Through debugging, optimization, and asset integration, I ensured a polished final product. Explore the game on itch.io, see it in action in the trailer, and check out my contributions in the VFX showreel and Miasma material video.

As Technical Lead for Forgive Me Not, I have driven the development of core systems to create an immersive and engaging gameplay experience. I engineered advanced AI systems to deliver realistic enemy behaviors, enhancing challenge and unpredictability. To deepen player immersion, I developed procedural camera animations that dynamically adapt to gameplay contexts, alongside responsive player movement mechanics for fluid navigation. I implemented a robust damage system to track health and influence gameplay outcomes, paired with visually striking VFX to enrich the game’s atmosphere. Character animations were designed for seamless transitions, ensuring natural movement flows. I also created an enemy spawning system tailored for level designers, enabling flexible and strategic enemy placement. To enhance exploration, I developed systems for interactive elements like weapons and doors, fostering deeper player engagement. Finally, I established compelling game objectives to guide player interaction, balancing challenge and progression. Check out the demonstration video and mechanic video to see the project in action. This work underscores my ability to lead technical development, integrating AI, gameplay mechanics, and visual fidelity into a cohesive experience.

The Miasma System, built in 4 weeks with C++ and Unreal Engine Blueprints, is an advanced particle tool for horror games, delivering dynamic, immersive environments. Recognized with 2nd place in the Tools/Systems category at GradEX 2025, it leverages AI-driven pathfinding and Environmental Query Systems for adaptive surface spreading, with features like surface traversal, dynamic type-switching, and optimized logic for scalability. Grotesque visuals, including pulsating veins and chromatic aberration, ensure immersion, with intuitive data tables for easy customization. Designed for scalability, it’s ideal for horror titles seeking organic, unpredictable threats. Discover its eerie potential with captivating Miasma types crafted for Miasma Ashlung: Demonstration Video.

Developed in Unreal Engine 5 using C++ and Blueprints, this road generation tool creates dynamic urban environments with procedural roads, intersections, and AI-driven traffic. I built a custom PosableMeshAsset class for performant boned mesh roads, implemented spline-based and physics-based car systems with A* pathfinding, and integrated procedural audio and VFX for immersive feedback. Features include customizable obstacles, nuanced driver profiles, and optimized performance via timers and reduced traces. Collaborating with a 3D modeler, I refined UV textures and enhanced intersection logic for realistic traffic flow. Explore the tool’s capabilities, from modular road networks to dynamic traffic behaviors, in this demonstration video.

Developed in three weeks using Unreal Engine 5, this dissertation project delivers an immersive boss battle system featuring adaptive AI and swarm intelligence. I designed a boss with a hierarchical state machine (HSM) and utility-based decision-making, managing three escalating stages with unique attacks—slam, throw, stomp, and meteor—each with dynamic cooldowns and animation montages. The boss spawns diverse enemy units (fast melee, slow melee, basic/advanced archers, healers, shields) using Environment Query Systems (EQS) to select optimal spawn locations based on pathfinding, proximity to the player, and tactical scoring. Swarm enemies employ a boids algorithm for realistic flocking, with configurable separation, alignment, and cohesion weights, enabling coordinated support for the boss while avoiding environmental collisions. I implemented a third-person player system with stance-based attacks (RMB + LMB), charged dashes, timed blocks, heals, and jumps, enhanced with VFX (e.g., magic aura) and sound effects with attenuation. The AI Perception System drives enemy detection via sight, sound, and damage, while a Blackboard shares critical data like player location. Optimization techniques, including culling off-screen agents, spatial partitioning, and multi-threading, ensure smooth performance. A custom UI displays boss and player health bars, animated ability cooldowns, a crosshair, and end screens for victory or defeat. Real-time player behavior analysis (health, distance, DPS) enables dynamic difficulty adjustment, creating a responsive challenge. Explore the system’s depth, from tactical AI to engaging player mechanics, in this demonstration video.

Developed in a high-pressure 32-hour sprint for the Ukie Game Jam 2025 by "The Weakest Game Group of All Time," Well of Ooze embodies the theme of "contrast." As the lead gameplay programmer, I drove the creation of core mechanics, procedural content generation (PCG), and game systems, while also bridging technical and artistic elements. The game follows a plucky slime escaping the shadowy depths of a procedurally generated well toward blinding light. I implemented responsive controls (right-click to aim, left-click to jump, WASD for mid-air nudges), a seed-based PCG algorithm for dynamic well layouts, and a data persistence system to save high scores and carry over level progression, enhancing replayability. I integrated audio, leveraging ElevenLabs’ generative AI for immersive sound effects, and crafted a rippling water material to elevate the game’s visual atmosphere. As the key liaison between tech and art, I ensured seamless integration of 3D models, UI, and VFX, delivering a polished experience. Key features include dual modes (Timed Normal and Endless), infinite replayability, and a cinematic finale. Explore the game on itch.io and see it in action in this demonstration video.

Over a 9-week Advanced AI module, I developed three distinct AI systems—Chess AI, Squad AI, and Long Term Planning AI—demonstrating proficiency in C++ programming, AI algorithms, and real-time game systems. Each project tackles unique challenges, showcasing intelligent, adaptive behaviors tailored to specific gameplay contexts.

I built a high-performance Chess AI using bitboards for efficient board state management and move generation. The AI leverages the Minimax algorithm with Alpha-Beta pruning to make strategic decisions, supported by a robust evaluation function that assesses material, piece positioning, king safety, pawn structure, and mobility. Optimizations like move ordering with Most Valuable Victim-Least Valuable Attacker heuristics and center control prioritization ensure a competitive and responsive opponent, ideal for challenging gameplay.

The Squad AI coordinates zombie agents in a dynamic game environment, employing a utility-based decision-making framework to prioritize actions such as pursuing the player, searching areas, assisting allies, or patrolling. A message-passing system enables agents to share critical information, like player sightings or distress calls, fostering emergent cooperative behaviors. Integrated with spatial awareness and pathfinding, the AI delivers immersive, intelligent enemy interactions, enhancing the game’s tension and strategy.

The Long Term Planning AI powers a resource management game where players assign workers to collect resources and perform tasks like constructing buildings. I designed an AI system that enables workers to autonomously organize and execute tasks efficiently using long-term planning strategies. The system optimizes task allocation based on resource availability and worker proximity, ensuring strategic efficiency. I implemented collision detection for precise environmental interactions, a countdown system for task timing, a robust input and rendering framework for seamless gameplay, and text metrics for dynamic UI displays, creating a cohesive and engaging player experience.

These projects highlight my ability to craft sophisticated AI systems that enhance gameplay through strategic decision-making, cooperative behaviors, and efficient task management. From competitive chess to dynamic squad coordination and resource-driven planning, this work underscores my expertise in AI development and real-time systems.

Over 4 weeks, I contributed to The Carver’s Lament, a 3D puzzle meditative platformer for mobile that blends unique mechanics with emotional storytelling. Players embody a world-class magical sculptor, chiseling mountains into monuments of famous figures like heroes and royals while uncovering the sculptor’s bittersweet journey. I designed and implemented innovative movement mechanics, allowing players to climb any mountain surface in 3D space by connecting visible points, and a sculpting system using tap, trace, and hold inputs to shape statues, reveal paths, and discover narrative items. These mechanics, paired with a 360-degree camera system, create an intuitive and immersive mobile experience. I also aim to develop systems for narrative integration, enabling players to collect items like books and pendants that reveal story fragments and grant permanent upgrades, such as increased sculpt range and travel speed. Environmental challenges like wind erosion and rumbles add depth, requiring players to adapt while learning patterns to improve. With a low-poly art style and soft piano score, the game will have 3-4 compact levels were rigorously tested on mobile to ensure smooth performance, focusing on physics-based movement and responsive UI. Watch the gameplay video to see the mechanics in action. The Carver’s Lament stands out for its emotional resonance and novel mechanics, offering a Journey-like experience with a crafting twist, tailored for mobile play.