Project Repositories
Portal Pemerintah Sumba Barat
Proyek portal Sumba Barat ini pada dasarnya adalah inisiatif pengembangan infrastruktur digital terpusat untuk domain sumbabaratkab.go.id. Secara fungsional, portal ini dibangun untuk menjadi sistem Satu Data Sumba Barat. Peran utamanya adalah mengintegrasikan berbagai data sektoral dari masing-masing instansi, memfasilitasi kelancaran proses verifikasi data di tingkat OPD (Organisasi Perangkat Daerah), dan menyediakan portal akses informasi publik yang lebih transparan bagi masyarakat. Dari sisi strategis, pengembangan portal ini merupakan langkah transformasi digital pemerintah daerah. Salah satu sasarannya adalah untuk meningkatkan kualitas layanan publik dan membantu optimalisasi Pendapatan Asli Daerah (PAD) melalui fitur pemetaan potensi pajak yang bisa dipantau secara real-time.
VIEW DETAIL
GKS Waikabubak
Website Gereja Kristen Waikabubak dengan fungsi kontrol dan pendataan jemaat terpadu seperti data nikah, baptis dan lain sebagainya
VIEW DETAIL
POS Offline First
Darul is a modern Point of Sale (POS) system designed to optimize printing business operations and workflows. Built with an offline-first approach, this application ensures that cashier transactions and service order recording can continue without obstacles, even if there is a disruption or disconnection of the internet connection. Main Features Offline-First Architecture & Auto Sync This system utilizes IndexedDB as a local storage medium on the client side. All cashier transaction data and new order input will be stored securely on the device directly. The synchronization logic embedded in it will automatically detect when the internet connection is stable again and upload the local data to the central server, ensuring that no data is lost (zero data loss). Service Tracking System (Service Tracking) Integrated tracking module that allows operators to monitor the entire progress of printing service orders in detail. Job status can be tracked and updated through various stages (example: Accepted -> Design Process -> Being Printed -> Finishing -> Ready to Take). This feature minimizes miscommunication and provides full transparency to work queues. Integrated Transaction Management Supports flexible transaction recording, both for selling physical goods (retail) and ordering service services in the same receipt or invoice. Responsive & Fast Interface UI/UX design optimized for speed of service at the cashier's desk, ensuring smooth navigation and instant application response time without the need to reload pages (page reload) thanks to the client-side data processing process. Project Objectives The main aim of developing Darul POS is to provide a cashier management solution that is resilient (resilient) to network infrastructure constraints. By combining the reliability of offline transactions and the ability to track the status of structured services, the system significantly improves the time efficiency, employee productivity and quality of customer service in the printing business.
VIEW DETAIL
Auto-Minutes: AI-Powered Meeting Transcription & Summarization System
An AI-driven web application engineered to significantly streamline the drafting of official meeting minutes and administrative records. This system automates the entire pipeline—from audio extraction (via URLs or local uploads) and highly accurate, GPU-accelerated local transcription (Speech-to-Text), to natural language processing (NLP) for intelligent structuring. The final output is a ready-to-use Microsoft Word document that details the meeting's chronology, speaker identification, conclusions, and action items, complete with the raw transcript appended for authentic archiving. Key Features: Multi-Source Audio Ingestion: Seamlessly extracts audio from YouTube URLs or accepts local media file uploads (MP4/MP3/WAV). Local GPU-Accelerated Transcription: Leverages local PyTorch-based AI models and NVIDIA CUDA architecture to deliver highly accurate, low-latency Indonesian speech-to-text processing. Intelligent Summarization: Utilizes a Large Language Model (LLM) with a massive context window to synthesize raw transcripts into chronologically structured, professional administrative formats. Automated Document Generation: Dynamically compiles the AI-generated summary and raw transcript into a downloadable Microsoft Word document (.docx). Asynchronous Processing: Built on an asynchronous backend architecture to ensure server responsiveness during heavy background AI computation. Tech Stack: Backend Framework: Python, FastAPI AI & Machine Learning: OpenAI Whisper (PyTorch / NVIDIA CUDA), Google Gemini API Media Processing: yt-dlp, FFmpeg Document Generator: python-docx Frontend Web: HTML5, Tailwind CSS, Vanilla JavaScript (Fetch API)
VIEW DETAIL
SIPANDIK (Conflict Monitoring and Detection Information System) – Regional Executive Command Center Infrastructure
SIPANDIK is a spatial intelligence and administrative infrastructure engineered specifically to facilitate the real-time monitoring, analysis, and prevention of conflict escalation at the local government level. Operating as an executive portal, the system acquires precise geolocation data from field agents, seamlessly transmits visual evidence to encrypted cloud storage, and automatically extracts reports into formally formatted electronic memorandums authenticated via QR Codes. Equipped with an interactive Command Center interface featuring satellite mapping and aggregate data visualization, this infrastructure automates the distribution of multi-tiered early warning notifications directly to executive leaders. This ensures an accelerated policy formulation and tactical decision-making process on the ground. DEMO: username: pimpinan password: rahasia123
VIEW DETAIL