Keep AI products confidential, intact and available
How AI products are secured in industry: the CIA triad, governance and privacy, cloud security for AI workloads, securing LLM applications, testing and assurance, and monitoring. Classic offensive and certification topics follow as an optional reference track.
7 stages · 15 topics
Confidentiality, integrity and availability, and how to model threats.
Up next · Stage 02Security has to be governed: policies, controls and the law.
Frameworks and regulation for data and AI systems.
Up next · Stage 03AI products run in the cloud, so secure the platform they run on.
Cloud security knowledge, including securing AI workloads and AI data.
Up next · Stage 04Then secure the AI application itself.
The attack surface that is new with LLMs and agents.
Up next · Stage 05Prove it: test the application and its APIs like an attacker would.
Web and API testing, vulnerability assessment and AI red teaming.
Up next · Stage 06No system is perfect. Detect and respond when something gets through.
SIEM, threat intelligence and incident handling.
Offensive security, OT/SCADA and certification bodies of knowledge covered in Ahmad's material.
Real projects built with the skills on this roadmap.
Projects coming soon
New portfolio work for this roadmap will appear here as it is published.