Early-Career Researcher in Machine Learning and Digital-Economy Research

Bintang ArigoKautsar Urumsah

An early-career researcher working at the intersection of machine learning, causal inference, and the economics of technology adoption.

I trained first in computer science and then in international economics. My undergraduate thesis at Universitas Gadjah Mada built a three-class opinion-mining system, implementing TF–IDF weighting and a Naive Bayes classifier from scratch. My master’s thesis at Corvinus University of Budapest used country-level panel data from the World Bank and V-Dem to examine how economic, educational and political conditions are associated with internet adoption. Between the two degrees I spent three years as a software engineer building data pipelines. I now want to bring causal and reliable machine learning to questions about how AI and other digital technologies spread across firms and countries — and with what consequences.

View My ResearchCV overview

Research interests

Questions I want to work on

Three of these areas rest on completed work; the others are developing directions for doctoral research. Each is marked accordingly.

Completed research

Master's ThesisCorvinus University of Budapest, 2024

Investigating the Multifaceted Factors Influencing Information and Communications Technology (ICT) Growth

A cross-country panel study that treats ICT growth, measured by the share of the population using the internet, as an outcome of social, economic and political conditions. Using country and year fixed-effects models on World Bank and V-Dem data, it finds that GDP per capita and FDI inflows are positively associated with internet use, that the association with educational inequality is positive in the global model but varies by region, and that the negative association with political competition is sensitive to model specification.

Read the full research summary

Applied work

Selected projects

Technical projects that show how I build, evaluate and question models, separate from my academic research.

  1. Multi-Class Sentiment Classification of Product Reviews Using Classical NLP Models

    How well can classical NLP pipelines predict the 1–5 star rating of a product review from its text alone, when the rating classes are heavily imbalanced?

    Independent project

    0.655Macro F1, Logistic Regression (held-out test set)

  2. Content Recommendation System for a Game Catalogue

    How can a catalogue of more than 460 games generate personalised recommendations for registered users quickly enough to serve them through a backend API?

    Professional role — Data Scientist, PT Dunia Sempurna Teknologi (December 2024 – present)

    < 1 msMean computation time

  3. Telecommunication Customer Churn Prediction

    Can a compact set of account features predict which telecom customers will churn, and how do resampling and tuning change the comparison between classifiers?

    Independent project

    0.75ROC-AUC, tuned SVM (test set)

All projects and case studies

Interdisciplinary trajectory

From computing to the economics of technology

Each step added a layer the next one relied on: programming and models first, then data engineering, then the economic questions those tools can address.

  1. 2015 – 2019

    Computer science

    Undergraduate training at Universitas Gadjah Mada, including machine learning, artificial intelligence and information retrieval.

  2. 2019

    Machine learning and NLP

    Undergraduate thesis: a three-class opinion-mining system with TF–IDF and Naive Bayes built from scratch.

  3. 2019 – 2022

    Software and data engineering

    Three years building ETL pipelines and integrating transactional data on an ERP platform.

  4. 2022 – 2025

    International economics

    MSc in International Economy and Business at Corvinus University of Budapest.

  5. 2023

    Panel-data analysis

    Co-authored course project on political stability using population-weighted fixed-effects models.

  6. 2024

    Technology-diffusion research

    Master’s thesis on the economic, educational and political correlates of internet adoption across countries.

  7. Next

    Causal and reliable ML for AI adoptionDeveloping

    A developing doctoral direction: credible, robust evidence on how AI is adopted and what it changes.

Let’s talk about research.

I welcome conversations about doctoral research opportunities, research collaboration, and work on machine learning and technology adoption. Email is the best way to reach me.

arigourumsah@gmail.comAbout me