Blog
Blog
Notes on machine learning, AI security and building things.
Introducing Folven: governed data pipelines for AI
Oct 27, 2026 · 5 min read
Companies want AI on their most sensitive data, and their auditors want proof of what happened to it. Folven is the platform I built to give them both: pipelines that mask personal data before any model sees it, and a tamper-evident record of every governed AI decision that anyone can check.
AI governanceData pipelinesProduct launch
Building Prooftrace: an AI tester that has to prove itself
Sep 28, 2026 · 11 min read
Acceptance criteria are still checked by hand, release after release. Here is how I built an AI tester that drives a real browser, proves each criterion, and hands back tests your team keeps, and the rule that keeps it honest: no proof, no pass.
AI agentsTestingPlaywright
Building Voice Auditor: the polite call that broke two rules
Sep 28, 2026 · 11 min read
A support call scored 99.9 out of 100 and still broke two company rules. Here is how I built Voice Auditor to check every call against written policy, what the first version got wrong, and how it keeps card numbers out of the database.
Speech AICompliancePrivacy
Detecting the Odd One Out: A Guide to One-Class Classification
Jan 8, 2025 · 6 min read
How one-class classifiers learn what normal looks like, when to use them, and a one-class SVM you can run in a few lines.
Machine learningAnomaly detection
How to Start with Data Science in 2020
Nov 8, 2020 · 4 min read
What data science is, the skills it takes, and the courses, books and podcasts I'd use to get started.
Data scienceCareers
What Will Your Plate Look Like in 2050?
Jan 13, 2020 · 5 min read
Population growth, soil degradation and a financing gap threaten the world's food supply, and what vertical farming, soil sensors and better financing could change.
TechnologyAgriculture