Projects

Engineered and deployed a responsive React and Tailwind CSS landing page that organizes textile products and tailoring services into a mobile-friendly customer experience.

Engineered a Retrieval-Augmented Generation (RAG) web agent with LangGraph orchestration and FAISS vector search to retrieve website context and generate targeted summaries through a deployed Streamlit interface.

Built an IndoBERT NLP pipeline for Shopee review sentiment analysis and exposed product-level sentiment and sales-performance insights through an interactive Streamlit dashboard.

Implemented a constraint-aware Genetic Algorithm and Flask application that optimizes resource and personnel assignments to generate conflict-free schedules.

Developed a Laravel and MySQL booking platform with availability tracking, reservation management, and administrative workflows for soccer-field operations.

Implemented a lightweight, real-time speech-to-text service with VOSK and FastAPI, optimizing inference for CPU deployment and integrating transcript summarization.

Designed a multi-role React and FastAPI system for farmers, administrators, and distributors to coordinate subsidized-fertilizer allocation and improve distribution transparency.

Developed a nutrition application with Gemini API-generated meal recommendations, personalized daily planning, and nutritional tracking across a React and Next.js interface.

Built an interactive React learning platform for Batak script, implementing experience-point and health-based gamification to structure learner progression and engagement.

Engineered an automated parking pipeline using YOLO for license-plate recognition and vehicle classification, integrated with a FastAPI service and PostgreSQL for persistent vehicle-event management.

Developed a Next.js AI-news aggregator that ingests updates from multiple public APIs and presents centralized, real-time industry coverage through a deployed web interface.

Built an interactive Power BI dashboard with product-tag, color, and name filters to analyze total sales profit, units sold by rating and category, and product-level performance.

Trained a Random Forest Regression model on Indonesian market and device-specification data to estimate smartphone prices, then delivered predictions through a deployed Streamlit application.

Implemented real-time vehicle detection, tracking, and counting with YOLO and ByteTrack, deploying the workflow in Streamlit for traffic-flow analysis.

Developed a YOLO cigarette-detection pipeline and Streamlit interface for automated image-based identification, review, and monitoring.

Designed and deployed a PHP, MySQL, and AdminLTE voting system for Hasanuddin University Faculty of Law Student Executive Board elections, hosting the application on a Biznet server with a Domainesia domain.

Implemented a StyleGAN generative modeling pipeline to synthesize realistic fish imagery for research experimentation and computer-vision data augmentation.

Engineered and trained a DCGAN image-synthesis pipeline to generate original anime character samples and evaluate adversarial-training behavior.

Architected and deployed a WordPress and Elementor website for the Data Science and AI Thematic Research Group, organizing research outputs and academic expertise for Prof. Armin Lawi.

Trained an attention-enhanced LSTM NLP model for cross-site scripting detection, achieving 98% training and validation accuracy, and deployed inference through a web application and browser extension.

Engineered an IoT cultivation system combining ESP8266 sensor integration, XGBoost growth prediction, and a Flask dashboard, achieving R² = 99.8% for fish length and weight estimates.

Benchmarked ResNet and YOLO road-crack models in PyTorch, using OpenCV preprocessing to compare classification performance and inference speed.

Engineered a semi-automatic irrigation controller with Arduino, Blynk remote operation, and timer-based watering logic to support more controlled water usage.

Built a mobile spice-classification workflow using YOLO and OpenCV, integrating deep-learning inference into an application for image-based identification.

Developed a YOLO and OpenCV computer-vision pipeline for palm-oil ripeness classification and automated object counting to support agricultural inspection.

Co-developed an Android application using ResNet50 and TensorFlow Lite for on-device gym-equipment classification, with Firebase, a Gemini API fitness assistant, and food-based nutritional tracking.

A web application that performs real-time facial emotion detection using TensorFlow.js and a pre-trained CNN model.

An interactive web app that estimates a person's age from their face in real-time using TensorFlow.js and DLDL-v2.

Implemented one-shot facial style transfer with JoJoGAN, PyTorch, and StyleGAN fine-tuning to generate Simpsons-style portraits while preserving identity characteristics.

Developed an interactive Tableau sales dashboard covering cost, revenue, profit, product and distributor performance, time-series trends, yearly comparisons, and geographic distribution.

Developed an Android eye-blink detection pipeline using MediaPipe facial landmarks and OpenCV for real-time landmark tracking and blink-state recognition.

A blog application with user authentication and CRUD functionality.

This project is a Drum Kit website built using HTML, CSS, and JavaScript. It allows users to interact with a virtual drum kit by pressing specific keys or clicking on drum elements.

A responsive landing page for a fictional startup called Tincat, built with HTML, CSS, and Bootstrap.

Short-form video edit delivered for Rodney The Finance Guy.

Short-form video edit delivered for DIAIM Marketing.

Short-form video edit delivered for DIAIM Marketing.

Short-form video edit delivered for The Benedetto Houston.

Short-form video edit delivered for The Benedetto Houston.

Short-form video edit delivered for EVIS Global - Smart Ergonomics Solution.