Data · AI · Jakarta, Indonesia
Building intelligent systems end-to-end — from raw data and machine learning pipelines to AI-powered applications backed by real government deployment experience.
About
Fresh graduate in Sharia Economics from UIN Syarif Hidayatullah Jakarta (GPA 3.40), combining quantitative foundations with hands-on AI and data engineering experience.
Currently working as Data Analyst at BPPK Kementerian Keuangan — building automation systems, dashboards, and an AI-powered budgeting assistant used in live government financial workflows.
Thesis applied Geographically Weighted Poisson Regression (GWPR) across Indonesian provinces, bridging academic econometrics with production data tools.
Every project approached end-to-end: from ingestion and feature engineering to modeling and user-facing interfaces. Not just analysis — systems that work.
Projects
AI-powered government budgeting assistant built with Streamlit and Gemini API. Retrieves financial data from 8 Excel sheets, performs automated SBM calculations, and generates structured budget tables. Deployed and actively used for internal financial workflows at BPPK Kemenkeu.
End-to-end analytics pipeline on 100k+ orders across 9 datasets. Includes data wrangling, feature engineering, RFM customer segmentation, revenue trend analysis, and geospatial mapping — delivered through an interactive multi-tab Streamlit dashboard.
Interactive Excel dashboard with PivotTables and slicers to monitor 300+ official travel assignments across 94 employees over a 2-month period. Used in production at BPPK Kemenkeu to support structured HR reporting and compliance tracking.
Power BI dashboard built from the same employee travel dataset — visualizing assignment distribution by team, monthly trends, and employee workload. Demonstrates cross-tool fluency: the same data source explored through both Excel and Power BI for different reporting needs.
Full end-to-end ML pipeline on 2,538 bank transactions — preprocessing, feature engineering, clustering, and classification to identify transaction patterns and surface analytical insights. Built with Scikit-learn from raw data to evaluated model.
Image classification model using TensorFlow and VGG16 transfer learning with data augmentation, callbacks, and ImageDataGenerator pipeline. Achieved over 97% accuracy on the Dicoding dataset — full training pipeline built from scratch.
Skills
Experience
Education
Thesis: GWPR Analysis of Internet Access, Education, Labor Force, and Sharia Financing Across Indonesian Provinces — applied Geographically Weighted Poisson Regression using R, GWmodel, spdep, and GeoDa to analyze spatial heterogeneity across 34 provinces.
Certifications