Data Science undergraduate at Universitas Negeri Surabaya (UNESA), specializing in machine learning engineering, distributed computing, and applied analytics. I build production-grade data systems, deploy predictive models, and translate complex datasets into actionable intelligence.

Transforming raw data into meaningful insights, one algorithm at a time.

Data insight and analytics hub for business intelligence and market trends

A platform delivering curated data narratives, analytics breakdowns, and market intelligence. Bridges the gap between raw data and business decision-making through accessible visual storytelling.

Depression risk detection system based on lifestyle indicators

A machine learning system that identifies depression risk through behavioral and lifestyle pattern analysis. Designed with clinical sensitivity and user privacy at its core, enabling early-stage mental health risk screening.

Circular economy platform to reduce food waste

A product design and platform concept addressing food waste through community redistribution and demand forecasting. Combines data-driven surplus prediction with an intuitive marketplace interface.

Custom Discord bot with music, administration, and community automation

A full-featured Discord bot supporting community servers with automated moderation, music streaming, event management, and custom command tooling. Built with a focus on reliability and extensibility.

Live BTC and ETH price tracker with 1-hour ahead predictions via 5-factor technical voting

A fully automated static web dashboard that fetches BTC and ETH hourly candlestick data from CryptoCompare, computes 5 technical indicators (Momentum 1H/3H, EMA crossover, RSI, Bollinger Band), and generates UP/DOWN/HOLD signals via majority vote. Predictions update every 15 minutes through a GitHub Actions scheduled workflow. AI narrative insights are generated via Groq (llama-3.1-8b-instant) and broadcast to a Discord channel as a formatted PNG card.

Architecture highlights:

• Cron-based GitHub Actions pipeline running at 3,18,33,48 * * * * with a 13-minute guard to prevent duplicate runs
• Signal logic: 5-factor vote where UP requires 4+ bullish votes; DOWN requires 2 or fewer; otherwise HOLD
• CairoSVG-rendered Discord notification cards with sparklines, indicator cells, and vote pip visualization
• Zero backend infrastructure: pure static site + data JSON committed by the workflow bot

Languages

Machine Learning and Data Science

Distributed Systems and Big Data

Web Development

Cloud and Infrastructure

Tooling

  

Statistical Modeling and Analysis
Machine Learning Engineering
Deep Learning Applications
Data Visualization and Storytelling
MLOps and Model Deployment