Data science

Understand the central limit theorem in plain language, why sample means look normal, and how it underpins confidence intervals.

I traced 1,327 URLs cited by the IANA Time Zone Database to see which historical sources still work and how many can be recovered from archives.

I traced 1,327 URLs cited by the IANA Time Zone Database to see which historical sources still work and how many can be recovered from archives.

Here’s why your RAG pipeline might have this incredibly common bug.

Inside the Core of Matter: My Experience at One of the World’s Leading Nuclear Research Centers

How to build web data pipelines 2.6x faster and 60% cheaper with Apache Arrow. Cut memory usage and file sizes by half for AI training.

No server. No cloud bill. Just DuckDB and 50M rows on a cheap $500 Acer laptop. Here's the honest truth on where local analytics work best.

How to use dbt’s layered SQL models and DuckDB’s embedded engine to clean, normalize, and transform raw web data into analyst-ready tables.

I got tired of grepping through JSON. So I built a local search index over live Google data using Python, Typesense, and Bright Data — here's how.

How to normalize, dedupe, and fuzzy-match records that refer to the same real-world entity in Python, without a database or any ML pipelines.