October 27, 2026
Stationnement, Not Parking: Building a Small Quebec French Scanner
Why sounding local matters on Canadian French websites. My experiment combining simple NLP, OQLF data and an LLM to find French gaps, and how it broke along the way.
AI in practice
Experiments in AI evaluation and real-world workflows, often through the lens of localization.
October 8, 2026
Every Rule Passed. The Translation Was a Page That Should Not Exist.
A short hotel title went into my AI translation pipeline and came out as an invented guest information page. Every check I had built said it was fine.
September 28, 2026
Context Engineering for Localization
A translation can be fluent and still be wrong. How a mistranslated post got me thinking about context engineering for localization, and why having the information isn't the same as using the right context.
September 14, 2026
Localization Already Had Retrieval. So What Does AI Add?
Where RAG and LLMs add value to the localization systems we already have.
September 9, 2026
The Reinforcing Loop Behind the AI Race
Three recent tech deals, Autodesk acquiring MaintainX, NVIDIA acquiring Hugging Face, and OpenAI launching GPT-6 Astra, reveal the same underlying pattern: reinforcing loops that make winning today easier tomorrow.
September 8, 2026
What I'm Learning About MT Quality Evaluation: From Scores to Decisions
How machine translation quality evaluation is evolving from traditional metrics toward LLM judges, MQM-style error diagnosis, publishability decisions, automated repair, and human-in-the-loop workflows.
August 29, 2026
What if AI Had Been Watching?
A look at whether AI powered computer vision could help prevent airport ground accidents, from a real safety check improvement at Toronto Pearson to what went wrong with LaGuardia's ASDE-X system in the March 2026 collision.