A hash chain in a few lines of R
How tamper-evidence works, and where it stops.
Field notes.
Short, practical write-ups for people putting R and AI into regulated work. Two are published, two are design notes and the rest are planned.
Write-ups.
Short, practical pieces. Planned ones are marked and will link to working code.
PublishedPlanned
Planned
Worked example: an audit trail and a replay check
A real run of regulog and reproducr, including a tamper test and a dropped row, with the output as it came out.
Read the example PublishedWhat the FDA and EMA AI principles ask of your R workflow
Each of the ten principles as a record you could show a reviewer, and how far the R packages get you.
Read the note Design noteMeasure the agent before you rely on it
How an evaluation of an AI coding agent could work, and what it cannot tell you. Pilot offer.
Read the note Design noteAI as the second programmer
How an AI agent could do independent programming QC, with every difference decided by a person. Pilot offer.
Read the noteReplaying an AI-assisted analysis
Locking the environment and comparing outputs.
Preparing R work for regulatory review?
Describe what your QA group or a reviewer would need to see, whether that is validation, a migration or the use of an AI assistant. Engagements are taken on in small numbers.