Projects

Rewop is an AI-guided platform for generating WebApps, built around structured backend workflows, persistent project state, and iterative generation loops.

Overview

Rewop is currently in active development toward its first public release, but a significant portion of its core platform functionality is already implemented. The backend supports user initialization, credit-based generation flows, project creation and retrieval, versioned bundle persistence, and AI-driven generate/update loops with validation and fallback handling.

What this project demonstrates

• AI-guided WebApp generation with structured multi-file output
• Frontend/backend workflows for project creation, retrieval, update, rename, and deletion
• Persistent user, project, and bundle state management with database-backed storage
• Cookie-based session handling and user initialization flows
• Versioned project bundles and saved generation history
• Credit-based usage logic for generation and update actions
• AI orchestration with retry handling, response validation, and fallback generation
• Production-minded API design, ownership checks, and error handling

Visit Rewop

WIP — v1 public release in progress.

FPS with hostile drones.

Overview

The FPS prototype takes place in a large urban test environment where the player must locate and eliminate hostile robotic drones. The level is intentionally sized beyond a small test arena to explore navigation, awareness, and sustained combat flow.

Drones operate in a patrol state (blue), and once the player is detected, they transition into an aggressive pursuit mode (red), accelerating toward the player and self-destructing on impact. Health and ammo pickups are distributed across the level, encouraging movement and route planning rather than static engagement.

What this project demonstrates

• First-person shooter controller with responsive movement and aiming
• Enemy state transitions (patrol → detection → pursuit → self-destruct)
• Dynamic threat behavior with visual state feedback (color-based)
• Explosive enemy interactions and area damage handling
• Timed gameplay loop with objective tracking (remaining enemies)
• Pickup systems for health and ammunition

Download prototype (Windows, ~600 MB)

This is a non-commercial prototype project.

Stealth NPC Prototype

Overview

This prototype runs in a small test level, but the NPC system is designed to scale to larger environments. It focuses on stealth-oriented behavior in Unreal Engine, with modular patrol, detection, and pursuit systems. Each NPC can be configured individually, including patrol timing and visibility cone radius, enabling varied guard behaviors across a level.

The system also supports coordinated response logic: when pursuit is triggered, nearby NPCs will be alerted, allowing for a more dynamic stealth pressure. This creates a foundation for expansion into larger stealth spaces with more systemic encounter design.

What this project demonstrates

• Stealth-oriented NPC patrol, detection, and pursuit logic
• Per-NPC configurability for patrol timing and visibility behavior
• Scalable AI setup suitable for larger stealth environments
• Coordinated multi-NPC response triggered by local events
• Player-driven avoidance, route planning, and spatial awareness
• Scope-controlled prototyping with room for systemic expansion

Download prototype (Windows, ~900 MB)

This is a non-commercial prototype project.