Blog

Whole-Page Optimization, One Year Later

November 5, 2026

At RecSys 2025 I had the only paper on building a whole recommendation page. At RecSys 2026 there were four. I read them as a map of where the problem moved, and of the part I am still working on.

Read More →

I Was Wrong About Git Worktrees, Twice

October 28, 2026

I used to dismiss Git worktrees as a waste of disk space I had no use for. Then uv made separate Python environments cheap, AI coding agents gave me a reason to work in parallel, and I went all in, until I found out that my brain is still single-threaded.

Read More →

Recommender Systems in the Age of Generative AI

May 7, 2026

At ML Prague 2026, Bogdan and I led a workshop on how recommender systems can be improved with generative models. We covered extending item metadata and transformer-based recommendation models, and planned to discuss generating a whole page.

Read More →

Forged in Rust, Spoken in Python

November 26, 2025

Python makes it easy to build and explore ideas, but performance-critical work often runs in native code. I found that Rust, PyO3, and Maturin made it practical to bring that speed into my whole-page recommender.

Read More →

Bringing the Carousel Experience to Recommender Systems

August 7, 2025

If you've ever opened your favorite app—whether for videos, music, shopping, or beyond—you’ve likely seen multiple rows—or "carousels"—of recommendations: “Keep Watching,” “Just for You,” “Hot Right Now,” and so on. Yet, much of the academic research out there still focuses on a single ranked list—a single column of suggestions.

Read More →