AI Red Teaming roadmap
Five stages · about three months at five hours a week · 22 resources
Test AI systems for weaknesses before attackers do.
Leads to roles such as AI Red Teamer, AI Security Engineer.
Create a free account to open the lessons and videos, tick off stages, save this roadmap and track your progress.
The stages
Stage 1: How AI systems fail
Models, data and the application around them.
about 2 weeks 4 lessons and videos inside
You will be able to
- Describe how language models are used in apps
- List common failure types
- Map an AI app's attack surface
Project
Draw the attack surface of a chatbot with tools.
Stage 2: Prompt attacks
Injection, jailbreaks and data leakage.
about 3 weeks 4 lessons and videos inside
You will be able to
More roadmaps to explore
- AWS CloudCore AWS services, set up safely on the free tier.
- DockerPackage, run and ship apps in containers.
- KubernetesRun containers reliably at scale.
- DevSecOps EngineerPut security checks into every step of building and shipping software.
- Cyber Security AnalystLearn to defend systems and recognise attacks.
- Network EngineerDesign and run the networks that everything else depends on.