Blog
Writing about AI, machine learning, data science, and operations research.
Solving the Subway Challenge by letting two LLMs run on autopilot
11 min readI model the NYC Subway Challenge as a time-expanded routing problem over the real MTA schedule. My current best validated route visits all 472 stations in 24:24:30.
What does it cost to see the world? Optimizing a flexible multi-stop flight search
10 min readFix only your home airport, list the cities you want to visit, and let a time-expanded traveling-salesman model pick the cheapest order and dates. We use it to price two bucket-list trips: the world's 10 most-visited cities and the New 7 Wonders.
Where does the New York Times fish for Spelling Bee puzzles?
8 min readIn honor of Scripps 2026, we analyze every one of the almost 3,000 honeycombs in the history of the NYT's Spelling Bee — and how they fare against all 7-letter words from Collins Scrabble Words. We conclude with a detour to another language, German.
A family like no other: studying the graph behind the Mathematics Genealogy Project
7 min readThis post is about two topics: first, as the title suggests, it is about studying the graph behind the Mathematics Genealogy Project and its many interesting — and, at times, surprising — properties (the "what"). But maybe more importantly, second, it is about "how" this post was written.
Are employees happier at small companies? I checked 10 million Glassdoor reviews to find out
5 min readWe analyze a Kaggle dataset with 10 million Glassdoor reviews to answer the question, "are employees happier at small companies?"