At work
Building the boring parts that determine whether digital commerce actually works.
Right now that means product data architecture, search quality, and the integration edges between commerce platforms and the systems that still run the business day to day.
In writing
I'm working through essays on product data economics, enterprise AI that has to survive procurement and governance, and the gap between demo-friendly software stories and operational reality in distribution.
Learning
I'm spending time on how LLMs can support classification, search, and internal knowledge work in large industrial catalogs where the data is inconsistent, the vocabulary is domain-specific, and the acceptable error rate is low.
Paying attention to
- Search relevance and synonym strategy in industrial catalogs
- How teams govern AI outputs inside quote-to-cash workflows
- The cost structure behind enrichment, taxonomy, and cleanup work
Last updated: April 2026