Arvi Rizqi
Fadhila
I build across the stack — from training ML models and deploying them as real apps, to crafting polished frontend interfaces. Fresh Informatics graduate, end-to-end builder.
4
ML Models Built
2
Frontend Projects
About Me
Building PredictiveProducts
I'm Arvi — a fresh Informatics graduate from Universitas AMIKOM with hands-on experience across machine learning, frontend development, and game development.
I don't just train models in notebooks — I architect full pipelines from data preprocessing and model training to production deployment. My frontend skills (React, Vue, Tailwind) let me build interfaces that make models actually usable. I also build interactive games and explore IoT, because real builders don't stop at one domain.
Currently open to
ML · Frontend Roles 🎯
🤖
4
ML Models Built
⚡
3
Frontend Projects
🎯
3.86
GPA
🎓
2026
Fresh Graduate
Featured Work
Projects
Spanning machine learning and frontend development — each project solves a real problem, start to finish.
Autism Classification
MobileViTv2 + TTA model for ASD classification from children's facial images
Early autism detection is often expensive, time-consuming, and requires professional clinical evaluation, making large-scale screening difficult.
Developed a computer vision-based classification system using MobileViTv2 combined with Test-Time Augmentation (TTA) to improve robustness across varying image conditions.
Achieved 90% classification accuracy. Successfully deployed as an interactive Streamlit web app for accessible ASD screening. Also implemented on the AutismeID website
⚡ If the Streamlit server breaks, simply run it again to restart.
Accuracy
~90%
Architecture
MobileViTv2 + TTA
Depression Prediction System
XGBoost-based mental health prediction from social media usage patterns
Mental health issues often go undetected due to lack of accessible early screening tools based on everyday user behavior.
Built a supervised machine learning model using XGBoost to predict depression risk based on user social media activity patterns.
Achieved over 95% prediction accuracy. Deployed as a public Streamlit web application for easy and real-time access.
⚡ If the Streamlit server breaks, simply run it again to restart.
Accuracy
~95%
Model
XGBoost Classifier
Plant Recommendation System
Random Forest model for plant recommendation based on soil nutrient conditions
Farmers and home gardeners often lack data-driven recommendations tailored to their soil conditions and nutrient availability.
Developed a Random Forest-based model to recommend suitable plants using soil nutrient inputs such as nitrogen, phosphorus, and potassium levels.
Achieved 99% accuracy. Deployed via Streamlit, enabling users to input soil conditions and receive instant plant recommendations. Also implemented on the j.tech website
⚡ If the Streamlit server breaks, simply run it again to restart.
Accuracy
~99%
Model
Random Forest
House Price Prediction
Gradient Boosting model for predicting house prices based on key property features
Accurate house price estimation is challenging due to multiple influencing factors such as location, size, and property characteristics.
Built a predictive model using Gradient Boosting trained on structured data with key property features to estimate house prices.
Achieved over 85% prediction performance. Deployed as a Streamlit app for real-time and user-friendly predictions.
⚡ If the Streamlit server breaks, simply run it again to restart.
Performance
R² ~ 0.90
Model
Gradient Boosting
J-Tech Crop Recommendation Platform
Professional company website built with React & Vite — AI integration planned
A sleek and modern company website for J-Tech, showcasing their AI-powered crop recommendation system. Built with React and Vite for optimal performance, the site features a responsive design and is currently live. Future roadmap includes integrating the crop recommendation ML model directly into the platform for real-time user access.
- Clean, modern design with a focus on user experience
- Responsive, mobile-first design using Tailwind CSS
- Production deployed and live at j-tech.arvrzq.my.id
- AI feature integration in development roadmap
🌐 AI-Powered Agriculture Platform
AutismeID
Autism awareness & early screening education platform for Indonesian parents
An empathetic web platform educating Indonesian parents and the public about Autism Spectrum Disorder. Features awareness content, early detection guidance, and integration with the MobileViTv2-based AI screening model — bridging research and real-world access.
- Built with React + GSAP for smooth, engaging animations
- Soft, accessible design palette for inclusivity
- Directly bridges the autism classification ML research
- Comprehensive ASD education content in Bahasa Indonesia
💙 Education Platform
Libris
Minimalist preloved book marketplace for affordable and sustainable reading
A clean and minimalist web platform for buying and selling preloved books. Libris promotes sustainable reading habits by giving books a second life, while making literature more accessible and affordable for everyone.
- Simple and minimalist UI for distraction-free browsing
- Focused on preloved books and sustainable consumption
- Fast and lightweight performance using Vite
- User-friendly listing and discovery experience
📚 Marketplace
Add-ons & Experiments
Another side of my work
Supporting work in UI/UX design and game development that complements my core technical skills.
Tech Stack
Skills & Tools
From ML pipelines to frontend UIs and game logic — the full toolkit of a multi-domain builder.
Machine Learning
Deep Learning
Data Analysis
Frontend Development
Deployment & Tools
✨ Multi-domain skillset — ML, frontend, and game development in one portfolio.
Background
Education & Credentials
Academic foundation and professional training that shaped my data & ML expertise.
Bachelor of Informatics — Universitas AMIKOM
Specializing in Computer Vision, Machine Learning, and Deep Learning. Final project: Medical Image Classification using CNN + Vision Transformer hybrid architecture.
Fullstack Java Bootcamp — Komdigi × Metrodata
Intensive enterprise software development program. Covered Java Spring Boot, RESTful API design, microservices, and production deployment workflows.
Vocational School — SMK SAKTI GEMOLONG
Software Engineering major. Built foundational programming skills and modern web development with HTML/CSS/JavaScript.
Get In Touch
Open to Data & ML Roles
I'm actively looking for opportunities to contribute to data-driven teams. Whether you're hiring, collaborating, or just want to talk ML — I'm reachable.
Contact Info
Location
Yogyakarta, Indonesia
arvrzqfdhla@gmail.com
arvi-rizqi-fadhila
GitHub
github.com/ArviRizqi