Project – Muhammad Dzaki Abrar
Autonomous Shape Sorting System (Smart Factory)
Prototipe otomasi industri yang mengintegrasikan Computer Vision (AI) dan IoT untuk mengklasifikasikan serta menyortir objek geometris secara otomatis di atas sabuk konveyor. Membandingkan YOLOv8 dan YOLOv11; memilih YOLOv11 karena 15% lebih sedikit parameter dan 25% lebih rendah GFLOPs. Mengembangkan logika non-blocking state-machine pada ESP32 untuk sinkronisasi presisi antara deteksi kamera dan gerakan pendorong mekanik, serta membangun dashboard interaktif untuk kontrol sensitivitas visi dan pemantauan statistik produksi.
Real-time Trash Detection
A real-time trash detection system built with YOLOv11 architecture that classifies different types of waste materials. The model achieves 85.12% mAP accuracy and can process video feeds in real-time for automated waste sorting applications.
NLP Hybrid Chatbot (RAG)
An intelligent chatbot leveraging Retrieval-Augmented Generation (RAG) with Llama 3 and Groq API. It retrieves relevant document context using FAISS vector search to provide accurate, context-aware responses.
AI Roadmap Adaptive E-Learning
An AI-powered educational platform that provides personalized, adaptive learning paths tailored to individual user needs and progress. Focuses on integrating AI logic into digital curricula to automatically determine the next optimal learning steps for each user.
PT. Schon Craft Indonesia
A professional web platform built for a local handicraft product company. This project involved a complete code rewrite to replace the old WordPress-based site. The primary focus was on improving site performance, page load speed, and scalability while maintaining a responsive and modern user experience.
Design Anything
A full-stack image editing platform that integrates various AI-powered features to enhance user creativity. Built with a Microservices architecture to separate core application functions, ensuring system stability and ease of future development.
Meet and Fight
A cross-platform mobile application developed as a final project for the Wireless Mobile Programming course. Focused on building a responsive user interface and implementing modern mobile application functionalities.
Student Management System
A desktop application designed for systematic student data management (CRUD operations). Implements Object-Oriented Programming (OOP) principles to maintain a modular and efficient code structure.
Industrial Escape: The Abandoned Factory
Trapped at night in an abandoned industrial factory controlled by a hostile presence, players must scavenge for components to manually repair the security system and unlock the main gate to escape. Features a fully playable 3D level implementing the 'Collect → Activate → Unlock' gameplay loop, utilizing Server/Client-side Lua scripts and a minimum of 10 custom 3D assets created in Blender.