Your company deployed AI. This is how you secure it before someone else tests it.
Securing AI Systems - Hands-on, for technical leaders. Most organisations wired a large language model into production in weeks and never asked how it could be attacked. This five-week, hands-on programme takes technical managers and architects through threat modelling, prompt injection and data exfiltration, securing retrieval and agent pipelines, data governance, and a structured AI red-team assessment. You finish with a findings report that turns "we deployed AI" into "we tested it".
Who this is for: Technical managers, security architects, and AI product owners.
The outcomes
What you will walk away with
Five weeks. Five artefacts. One file of proof.
Threat Model for a Sample App
A complete threat model for a working LLM application, built against the OWASP LLM Top 10.
Lab Report
Documented findings from the prompt-injection and data-exfiltration lab, with the mitigations that held.
Secure Architecture Diagram
A reference architecture for RAG, agents and tool use that your engineers can build against.
Data Handling Policy
Access control and data governance for AI systems, written as policy your organisation can adopt.
Red-Team Findings Report
The capstone: a red-team findings report on a working AI system, evidence that you can secure what you ship.
Then the certificate that proves it all
Every artefact above rolls up into a verifiable credential with its own serial number and public verification page. Your board can check it. Anyone can.
The syllabus
The programme, week by week
The OWASP Top 10 for LLMs, applied to a real app.
- LLM Threat Modelling30 minWatch free
You build:Threat Model for a Sample App
Before you commit
Questions, answered
Your cohort begins 9 September 2026.
CardBank transferaccepted
