A blog about nothing (and baseball).
I’m Sam Sharpe, a Staff Engineer in Applied AI Research at Capital One, building foundation models for customer behavior. Before that I built metrics for Statcast at MLB. Here I write long-form analyses of baseball and the occasional question about nothing in particular.
Research
Google Scholar ↗
2026
Dynamic Tokenization via Reinforcement Patching: End-to-end Training and Zero-shot Transfer
NeurIPS 2026
2025
Understanding the Effect of using Semantically Meaningful Tokens for Visual Representation Learning
CVPR 2025 · Workshops
2024
A Simple Baseline for Predicting Events with Auto-Regressive Tabular Transformers
Preprint · arXiv
Writing elsewhere
Capital One Tech
Dynamic Customer Embeddings & Understanding Customer Intent
How sequential recommendation and representation learning can be leveraged to model customer behavior.
Capital One Tech
Machine Learning for Temporal Data in Finance
How machine learning methods are evolving to address the complex temporal and sequential elements of consumer financial data.
MLB Technology
MLB Pitch Classification
A deep dive into recent enhancements of MLB's pitch classification system.
MLB Technology
An Introduction to Expected Weighted On-Base Average (xwOBA)
An introduction to new baseball metrics derived from tracking data.
All writing
| Injuries in Baseball: How (Self-)Exciting? | Baseball | 13 min | |
| A/B Testing the Shift | Baseball | 14 min | |
| 2019 MLB MVP Predictions: Declining Focus on Batting Average | Baseball | 5 min | |
| All Betts Are on Yelich | Baseball | 4 min | |
| Making Baseball Slow Again | Baseball | 9 min | |
| Thinking Like an MLB MVP Voter | Baseball | 11 min | |
| Who Are the Best Romcom Actors? | About nothing | 10 min | |
| Stealing Bases and Splitting the Rewards | Baseball | 13 min | |
| Profiting off the Nationals Presidents Race Puppet Masters | About nothing | 10 min |