I’m Bart Amin, a data scientist working in industry, where I apply statistical modeling and machine learning to solve real-world problems. My work focuses on translating complex business challenges into well-defined analytical questions and delivering insights that drive practical decision-making.
I hold a Master’s degree in Methodology & Statistics (Behavioural Data Science) from the University of Amsterdam, where I developed a strong foundation in (Bayesian) statistical inference, machine learning, forecasting, and programming. During my time as a visiting student at Harvard University, I further strengthened my understanding of research methodology under the guidance of leading experts in the field.
I believe an effective data scientist combines two key strengths: the ability to rigorously frame problems using the right data and methods, and the ability to communicate results in a way that leads to clear, actionable outcomes.
CDRAG CDRAG (Clustered Dynamic Retrieval-Augmented Generation): LLM-selected cluster retrieval for RAG — outperforms standard cosine retrieval on legal QAStandard RAG systems retrieve the top-K most
Confidence intervals (CIs) are a powerful and elegant way to quantify uncertainty in estimates. Instead of providing a single-point estimate—such as an average revenue increase
Clustering is a popular (unsupervised) technique in data science, widely used for discovering hidden patterns in data without having prior understanding of what such patterns
*Disclaimer: The values and predictions presented in this blog post are entirely random and do not reflect actual findings. Additionally, this project was not conducted on behalf