Medical Device Software V&V Training Course + SW Documentation Kit Gold

Where an AI software file comes apart

A team documents IEC 62304 properly and treats the model as an implementation detail, so the file describes how the code was written and says nothing about how the model was trained. Somewhere else, the training data is described in a slide deck that was never version controlled, and by the time a reviewer asks where the data came from nobody can reconstruct it. The intended purpose is written for the software, and the finished product — the thing the user installs and runs — is never validated as a product at all.

And the one that costs the most time: every planned retraining is treated as a change to be assessed later, so there is no pre-determined change control plan, and the first update after certification turns into a change notification.

None of these are hard problems. They are the consequence of documenting a learning device as if it were ordinary software with a model bolted on.

• 4 modules
• 2h 30 of video
• 23 templates
• Word and Excel
• Self-paced
• Certificate on completion
• Access does not expire

Scope

What the bundle covers

The course covers IEC 62304 and IEC 82304-1: how the software is built, and whether the finished product does what it is meant to do. The Gold kit adds what comes on top of that when the device learns — the AI/ML lifecycle documents, data governance, model validation and the post-market monitoring that goes with them.

Contents

Four modules, then 23 templates in three coordinated modules

The course comes first: it explains the reasoning and the sequence. The templates are the structure that reasoning produces, already written and cross-referenced.

Module 01 holds the development plan, architecture and detailed design, software requirements specification, traceability matrix, SOUP list and evaluation record, V&V plan and report, release and configuration management record, and the problem resolution and anomaly log. Module 02 holds the AI/ML lifecycle plan, data governance and bias assessment, model validation, AI risk management to AAMI CR34971, the pre-determined change control plan, human oversight and transparency, post-market performance monitoring, the AI-augmented clinical evaluation plan and report, and the governing SOP. Module 03 holds the cybersecurity risk assessment and threat model, the cybersecurity management plan and the vulnerability disclosure policy.

Audience

Who the bundle is for

• Regulatory affairs and quality managers at manufacturers of AI-enabled software devices
• PRRC and quality managers who have to sign off a software file they did not write
• Software leads and technical founders building their first IEC 62304 file
• Consultants who need a defensible starting structure rather than a blank page
• Internal auditors preparing to assess a software and AI file

It assumes you already run a quality management system and know what your device does. It does not teach software engineering, and it does not teach machine learning.

What is included

What you get when you buy

• 4 video modules, 2h 30 in total, with module 00 free to watch before you buy
• A downloadable transcript and the slide handouts for every module
• A test at the end of each module, and a final test
• The software file gap analysis workbook
• 23 editable templates in Word and Excel, in three coordinated modules
• A certificate with a verifiable code
• Every future revision of the course and of the templates, at no further cost

Questions

Before you buy