Rahfi's | Projects.
I have worked on various projects, both personal and collaborative, that showcase my skills and creativity. Here are some of my notable projects that I am proud to share.
rahfi.pro
A personal portfolio website showcasing my experiences, educations, skills, projects, achievements and writings. Built with Next.js and Tailwind CSS.
AEGIS: AI Engine for Grounded Inspection System - Accenture Use Case
AEGIS (AI Engine for Grounded Inspection System) is an intelligent predictive-maintenance platform designed to transform traditional reactive maintenance into a data-driven, adaptive, and automated system. The project leverages machine learning, real-time sensor analytics, and Agentic AI to detect anomalies, predict equipment failures, and recommend optimal maintenance schedules across industrial sectors such as energy, mining, and manufacturing. By closing the gap between abundant sensor data and actionable insights, AEGIS enables industries to reduce downtime, enhance safety, and improve operational sustainability.
SATRIA: System for Analyzing Total Risk of Agent & Group (Sistem Analisis Terpadu Risiko Agen & Grup)
SATRIA is an AI-based early warning system designed specifically for microfinance risk operations. It integrates field agent performance analysis and borrower group health to detect potential default risks early.
SATRIA: System for Analyzing Total Risk of Agent & Group (Sistem Analisis Terpadu Risiko Agen & Grup) - Prototype
SATRIA is an AI-based early warning system designed specifically for microfinance risk operations. It integrates field agent performance analysis and borrower group health to detect potential default risks early.
Stacked Pretrained Models Pipeline for Javanese Scripts (Aksara Jawa) Handwritten Character Recognition and OCR
A stacked pretrained models pipeline combining 3 pretraned models (MobileNetV2, EfficientNetB0, ResNet50) with ANN as meta learner improved with Tree-structured Parzen Estimator (TPE) for hyperparameter optimization to enhance handwritten character recognition and OCR accuracy on Javanese Scripts (Aksara Jawa) dataset.
Hybrid CNN Vision Transformer (ViT) Model for Enhanced Image Classification
A hybrid model combining Convolutional Neural Networks (CNN) and Vision Transformer (ViT) architectures to improve image classification performance by leveraging the strengths of both approaches with embedding patch.
Unified ConvNeXt Pipeline for Image Classification, Object Detection, and OCR
A unified pipeline for ConvNeXt models to perform image classification, object detection, and OCR tasks efficiently.
The Assistant - Gemma 3: 4B Powered Chatbot Assistant
The Assistant is a Gemma 3: 4B Powered chatbot assistant designed to help you explain and generate knowledege faster and more efficiently. Provide the things you and he's ready to help you.
Codepilot
Codepilot is an Qwen 2.5-Coder: 7B Powered coding assistant designed to help developers write code faster and more efficiently. It provides for code generation, fixing code, refactoring and explaining code, making it easier for developers to focus on their work and improve productivity.
An MLOps Approach for Heart Failure Prediction Using MLP and Bayesian Optimization via Tree-Structured Parzen Estimator (TPE)
The goal of this MLOps project is to build and deploy a robust heart failure prediction system using machine learning model. The project will use an MLOps lifecycle to automate the training, optimization, and deployment of the model.
Optimizing Stroke Risk Prediction Using Artificial Neural Networks with Feature Selection Based on Binary Particle Swarm Optimization
This project aims to develop an optimized stroke risk prediction model using Artificial Neural Networks (ANN) combined with feature selection based on Binary Particle Swarm Optimization (BPSO). The goal is to enhance the accuracy and efficiency of stroke risk assessment by identifying the most relevant features and fine-tuning the ANN model.
Asah - Course Submissions, Code Resources, and Docs
Repository showcasing all work from the Asah - Machine Learning, including course submissions, code resources, and documentation.
inTrain.ai: Mobile AI-Powered Application for Interview Preparation and Upskilling
inTrain.ai is an AI-powered mobile application designed as a comprehensive solution for professional career planning and development. The platform integrates Large Language Model (LLM) technology to provide an AI HR Chatbot (Voice talk AI interview in newer version!), allowing users to engage in interactive mock interview simulations across various difficulty levels and job scenarios to receive real-time feedback and evaluation. Beyond interview practice, inTrain.ai facilitates holistic job preparation through a CV Analyzer for ATS-standard document optimization, personalized Job Roadmaps, and a curated course catalog for upskilling, as well as a Mentorship Matchmaking feature that connects users directly with industry practitioners. (Coming soon and stay tuned as full production grade apps!)
EzTrip: AI-Powered Travel Companion & Automatic Itinerary Planner
By leveraging AI-powered recommendation engines, EzTrip offers personalized suggestions based on users' interests and travel goals, streamlining the process of destination discovery, booking, and expense management. With an integrated expense tracker and generated itineraries, users can monitor their travel budgets in real-time, making it easier to plan and control spending throughout their journey. Our team is focused on enhancing Indonesia’s tourism experience by making local insights accessible to global tourists and supporting sustainable tourism growth. This project seeks to solve real challenges in destination planning for foreign travelers by simplifying their access to curated travel information & budgeting.
Bangkit Academy 2024 Batch 2 - Course Submissions, Code Resources, and Docs
Repository showcasing all work from the Bangkit Academy 2024 Batch 2 - Cloud Computing Cohort, including course submissions, code resources, and documentation.
Dicoding ML Engineer Path - Course Submissions, Code Resources, and Docs
Repository showcasing all work from the Dicoding ML Engineer Path self-learning subscription, including course submissions, code resources, and documentation.
dotfiles
Personalized dotfiles repository to streamline and enhance my development environment across multiple systems. It includes configurations for shell, editor, and various tools to ensure a consistent and efficient workflow.
Commit Wave
Automated GitHub Commits Bot: This bot will maintain your github commit history using datetime range and will commit randomly based on your input
If you want to see more of my projects check out my GitHub or press the button on the navbar.
Rahfi's | Certifications.
I have obtained various certifications that demonstrate my expertise and commitment to continuous learning. Here are some of the certifications that I am proud to showcase.
Certifications
Learning Certificates
Membangun Proyek Deep Learning Tingkat Mahir
Dicoding Indonesia
Belajar Matematika untuk Data Science
Dicoding Indonesia
Membangun Sistem Machine Learning
Dicoding Indonesia
If you want to see more of my certificates check out my LinkedIn or press the button on the navbar.
