Showing 42 projects
SIRESITA | Destination Recommendation System for North Sumatra
Capstone project focusing on building a tourism recommendation system for North Sumatra using two core machine learning techniques: Content-Based Filtering (CBF) and Collaborative Filtering (CF). The models were developed from scratch using TensorFlow and Scikit-learn, without relying on pretrained APIs or external services.
SIPENDEKAR | AI-Based Road Damage Detection & Prediction
An award-nominated AI system for detecting and predicting road damage using multitask learning with a custom Swin Transformer model. The model outputs multiple predictions including damage type, size, repair time, and material needs from a single image input. I led the machine learning pipeline, including data enrichment, model development, and deployment via Flask API.
Fruit & Vegetable Classifier + Nutrition Info WebApp
An image classification web app built as a final project for a machine learning course. It uses an ensemble of ResNet and Swin Transformer models to classify 32 classes of fruits and vegetables, providing nutritional facts such as calories, fiber, and vitamins. The ensemble uses SVM as a meta-learner to improve prediction accuracy.
Manggrow.id | AI Plant Assistant & Internship Management
A comprehensive plant monitoring and internship management web application integrated with an intelligent AI assistant. The AI uses a RAG system to answer specific gardening and internship questions based on a curated knowledge base. The backend leverages n8n for orchestration, connecting Supabase, and integrates AI-powered conversational agents.
Enterprise Asset Classification | Power Platform + AI
An end-to-end asset management solution for corporate environments. It combines Microsoft Power Apps for the user interface and Power Automate for approval workflows. The core intelligence is powered by n8n and LLMs to automatically classify asset categories based on descriptions, storing data securely in SharePoint.
YOLOv11 & DeepSORT Speed Estimation | Real-Time Traffic Analytics
A high-performance computer vision application for analyzing traffic flow in real-time. Combines state-of-the-art YOLOv11 object detection with DeepSORT tracking to estimate vehicle speeds from video feeds. Built from scratch with custom PyTorch implementations of YOLOv11 architecture and DeepSORT tracking algorithms, demonstrating deep understanding of modern CV pipelines.
InsightSphere | Enterprise Finance AI RAG System
An enterprise-grade financial analysis platform powered by RAG (Retrieval-Augmented Generation) technology. Enables financial analysts and consultants to instantly query complex documents (PDFs, Excel, Reports) and receive accurate, context-aware insights. The system supports multi-persona AI responses with ultra-fast inference powered by Groq and ChromaDB vector storage.
DERMA-DFU.ID | AI-Powered Diabetic Foot Ulcer Triage System
A comprehensive medical AI web application for intelligent screening and triage of diabetic foot ulcers. Leverages state-of-the-art deep learning models for infection detection, ischemia assessment, and automated wound measurement, providing evidence-based clinical decision support for rural healthcare workers and patients.
River Gauge AI | Intelligent Water Level Monitoring System
A production-grade computer vision system for automated hydrological monitoring. Leverages YOLOv8 object detection to identify and read water gauge markings from RTSP video streams or uploaded footage. Features intelligent waterline detection using texture analysis and robust tick fitting algorithms, with real-time alerting for critical water levels.
Product Demand Forecasting | LSTM Time Series Prediction
An advanced time series forecasting project that predicts weekly product demand across multiple sites using deep learning. The model employs LSTM neural networks built with TensorFlow/Keras to analyze historical sales data and generate accurate demand predictions for inventory optimization.
End-to-End MLOps Workflow
A demonstration of a complete Machine Learning Operations (MLOps) pipeline. It integrates MLflow for experiment tracking, DagsHub for collaboration, Docker for containerization, and Grafana/Prometheus for system monitoring.
Customer Segmentation using RFM Analysis | K-Means Clustering for Retail Marketing
An advanced customer segmentation project utilizing RFM (Recency, Frequency, Monetary) analysis combined with K-Means clustering to identify distinct customer groups for targeted retail marketing strategies. Built entirely with Python, Pandas, and Scikit-learn, this analysis processes over 1M transactions to segment 5,000+ customers into actionable marketing personas.
Face Recognition Attendance System | Anti-Spoofing Security
An enterprise-grade automated attendance platform designed for universities and corporations. Leverages AI-powered face recognition with advanced anti-spoofing technology to provide secure, contactless, and instant verification. The system combines YOLOv11 for face detection and InsightFace for recognition, featuring active liveness detection to prevent photo and video replay attacks.
Static QnA Chatbot
An NLP-based chatbot that answers user questions from a predefined QnA document. It uses text preprocessing, TF-IDF, and cosine similarity to retrieve the most relevant answers. The frontend is built with ReactJS + TailwindCSS and the backend uses NodeJS.
Chat With Your Documents | RAG System
A Retrieval-Augmented Generation (RAG) application that allows users to upload documents and chat with their content. It uses LangChain, Pinecone, and Groq LLaMA API for semantic retrieval, with a Streamlit UI and FastAPI backend.
Nail Virtual Try-On | AI-Powered Beauty Tech Platform
An enterprise-grade virtual try-on platform for nail polish powered by advanced AI segmentation. Enables users to instantly preview thousands of nail polish colors and finishes on their own hands through real-time computer vision and WebGL rendering. Features 95%+ accuracy nail detection using TensorFlow.js and MediaPipe Hands, with complete client-side processing for maximum privacy.
Waste Classification | Organic vs Recycle
An image classification project that classifies waste into Organic or Recycle categories. The model is built using TensorFlow and Keras, while Streamlit is used to display predictions and probability charts.
Playstore App Comment Sentiment Analysis
A text classification model for categorizing Playstore app reviews into Positive, Negative, and Neutral sentiments. The project is developed in Google Colab using TensorFlow and an LSTM-based architecture.
WEB.DERMA-DFU.ID | Healthcare E-Commerce & Partner Platform
A comprehensive healthcare e-commerce platform for diabetic wound care products in Indonesia. Features multi-language support, product marketplace with shopping cart, partner registration system, educational hub, and AI-powered chatbot for customer service using n8n RAG automation.
Multimodal Emotion Classification | Building and Comparing Unimodal vs Multimodal Approach
A comprehensive deep learning project comparing the performance of unimodal (text-only, image-only) versus multimodal (text + image) models for emotion classification. Features advanced techniques including OCR fusion, feature gating, and contrastive learning to classify 7 emotion categories with high accuracy.
Concrete Defect Detection | Enterprise Structural Health Monitoring
An enterprise-grade AI platform for automated civil infrastructure inspection. Leverages advanced Computer Vision to detect, classify, and assess the severity of 5 types of structural defects in concrete columns. Designed for civil engineers and inspectors to transform manual visual inspections into digital, auditable, and quantifiable workflows.
StockAI - IDX Prediction Platform | AI-Powered Stock Market Analysis
An enterprise-grade financial analytics platform for predicting stock movements on the Indonesia Stock Exchange (IDX). Combines Machine Learning with Bandar Analysis (Broker Accumulation) to identify high-probability trading signals. Designed for traders and analysts to transform raw stock data into actionable insights through a modern, responsive dashboard.
cGAN & VGG16 for Scalp Disorder Classification
A multi-class image classification system combining Conditional GAN for synthetic data generation and VGG16-style architecture for medical image classification of scalp disorders. Built from scratch using TensorFlow/Keras and PyTorch for data augmentation through generative modeling.
Enterprise PPE Detection VMS | AI Video Management SaaS Platform
A cloud-connected enterprise Video Management System for industrial health & safety compliance. It extends an on-premise YOLOv8 PPE-detection desktop app with a full B2B SaaS layer: a Next.js dashboard, hardware-locked licensing, live multi-camera WebRTC streaming, and automated incident management. Built for a petroleum services company to monitor PPE compliance across multiple sites.
Integrated Business Automation Suite | POS, Warehouse & AI WhatsApp Agents
A multi-module automation ecosystem built for a growing franchise business. The suite combines a web-based Point-of-Sale system for up to 500 outlets, a real-time warehouse management dashboard, an admin profit-and-loss reporting panel, and two AI-powered WhatsApp chatbots for automated business intelligence and operational coordination.
Mokapos to WhatsApp Automation | Daily Sales Reporting & Customer Campaigns
An n8n-based automation suite that integrates the Mokapos POS API with WhatsApp to eliminate manual daily reporting for a bakery franchise. Every night it fetches per-outlet revenue from the Mokapos API, formats a consolidated report, backs it up to Google Sheets, and delivers it to the owner via WhatsApp — fully unattended. Two extra workflows handle birthday greetings and promotional blasts.
AI Patient Journey Automation | Medical Tourism Clinic Platform
A large-scale n8n automation platform that orchestrates the entire patient acquisition and clinical pipeline for a cosmetic surgery clinic. It handles inbound WhatsApp leads from Meta Ads, drives patients through a multi-stage AI-managed funnel (qualification, consultation, medical intake, lab review, treatment proposal, payment, surgery scheduling, post-op recovery), and uses LangChain AI agents with a pgvector RAG knowledge base for bilingual communication.
Car Subscription & Long-Term Rental Website | Australia
A full-stack marketing and enquiry website for an Australian car subscription and long-term rental business. Built with Next.js 14, Supabase, and Tailwind CSS, it features a filterable vehicle catalogue, detailed car pages, a customer enquiry form with email notifications, and a complete admin panel for managing cars, enquiries, testimonials, FAQs, and SEO content.
Scaffolding Rental Management System | Web App Replacing Excel
A cloud-based web application built for a scaffolding equipment rental company to replace an entirely manual, Excel-based workflow. The system automates stock tracking across a central warehouse and multiple active construction projects, calculates monthly and prorated daily invoices, and enforces a two-role approval workflow before any stock movement takes effect.
Kindergarten WhatsApp Chatbot | Admissions & Info Bot
An n8n-based static chatbot deployed on WhatsApp to automate inbound enquiries for a private kindergarten. It presents a structured menu that guides parents through school profiles, new-student enrollment procedures, curriculum details, fee breakdowns, seat availability, and contact information — eliminating repetitive manual responses from staff.
Car Parts Catalog & Workshop Booking Website | Australia
A full-stack Next.js website for an Australian auto parts and workshop business, providing a public-facing parts catalog with car-compatibility filtering, a workshop service booking form, and a self-service admin panel so the client can manage products, bookings, and inquiries without developer assistance.
Cooperative Automation Platform | WhatsApp AI Agent + Member Wallet
A large-scale automation system for a savings cooperative and travel program, combining an AI-powered WhatsApp chatbot for lead nurturing and member self-service with a Xendit-integrated wallet for deposits, withdrawals, and recurring savings plans. Conversations are routed between two specialized AI agents depending on whether the user is a prospective or an active member.
Multi-Channel CS Automation | WhatsApp & TikTok AI for a Window Furnishings Retailer
A multi-channel AI customer service automation for a window blinds retailer. It handles inbound WhatsApp inquiries through an n8n workflow powered by a Claude-based AI agent, and covers the TikTok Shop Message Center with keyword replies, AI chat prompts, and suggested questions. The system qualifies leads, calculates price estimates, generates PDF quotations and DP invoices, and escalates to human agents when needed.
Odoo × Shopee Integration | Automated E-Commerce Sync
A full bidirectional integration between Odoo 17 (ERP) and the Shopee Open Platform for a fashion retail business. The system automates product and stock synchronization, order ingestion, and variant management so the merchant never manages two systems manually. Built entirely on n8n workflows deployed alongside Odoo via Docker.
Spray Injector Nozzle AI Inspector | Computer Vision Quality Control
A Computer Vision system built for a quality-control team to automatically measure four geometric spray attributes and classify the condition of fuel injector nozzles (Normal / Repair / Replace) in real time. It replaces subjective manual inspection with a consistent, AI-driven pipeline running on a Raspberry Pi 5, eliminating operator variability on the test bench.
AI Camera Activity Recognition System | Multi-Object & Human Activity Detection
A real-time AI camera system that detects multiple object categories (humans, animals, heavy equipment, everyday items) and recognizes a wide spectrum of human activities and facial expressions using YOLO-based pose estimation and a temporal action classifier. The desktop application targets use cases such as homecare monitoring and restaurant staff supervision, running fully offline with a SQLite activity log.
AI Sales Automation Agent | Smart Follow-Up, Lead Scoring & CRM
An intelligent sales automation system built on n8n for a vehicle paint-protection film (PPF) and vinyl-wrap installer. The system extends an existing Instagram DM AI chatbot with a context-aware follow-up engine, automated lead classification (HOT / WARM / COLD), a Google Sheets CRM that captures customer data and financials, and automatic invoice dispatch — all without manual intervention from the sales team.
WhatsApp AI Sales Chatbot + Odoo ERP | End-to-End Order Automation
A full-stack sales automation solution combining a WhatsApp AI chatbot (built on n8n) with a freshly implemented Odoo ERP backend for a medical equipment and supplies distributor. The AI agent handles customer identification, product catalog browsing, order creation, proforma invoice dispatch, and admin escalation — turning WhatsApp into a low-effort sales channel backed by a real ERP.
JerseyGen | AI-Powered Football Jersey Design Studio
A Next.js web application that lets users design custom football and futsal jerseys with a real-time 3D preview, then uses an AI image-generation API to produce a photorealistic render of a person wearing the customized jersey. Users can design patterns, colors, text, and logos, optionally upload their own face photo, and download or print the final design sheet.
MDCNN News Text Classifier | Multi-Teacher Knowledge Distillation
A research replication project implementing the MDCNN (Multi-Teacher Distillation-Based CNN) architecture from a 2025 IEEE Access paper. The model combines BERT and BiGRU as teacher models to train a lightweight TextCNN student via knowledge distillation, achieving near-BERT accuracy on the AG News classification benchmark at a fraction of the model size.
Ad Performance Monitor & AI Advisor | Multi-Channel ROAS Automation
A self-hosted automation system that pulls daily advertising data from Meta Ads, Shopee, and TikTok Shop, calculates ROAS and key metrics, and delivers AI-generated optimization recommendations to a Telegram group. It combines n8n for workflow orchestration with an AI agent for conversational advisory and long-term memory.
Fuel Supply Chain AI | Demand Forecasting & Distribution Optimizer
A Streamlit-based AI dashboard built for a fuel distribution terminal, covering demand forecasting for 135 retail stations over a 1–10 year horizon and a multi-agent AI architecture prototype. The system uses XGBoost and Prophet models with population and vehicle-growth factors, displays an interactive map, and includes a conversational LangChain AI assistant for supply chain queries.