My Portfolio

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.