AI Safety Engineering
Implement guardrails, red-team prompts, and build safer AI applications.
0%
Confidence · no data yet
After this skill you can…
- Implement input and output guardrails
- Red-team a deployed LLM application
- Use Llama Guard or NeMo Guardrails
Prerequisites
Watch (10 videos)
I Broke Threads
→ Design and test secure systems to prevent crashes and errors→ Develop and implement safety protocols for app development→ Analyze and mitigate potential security risks
From Assistant to Adversary: When Agentic AI Becomes an Insider Threat
→ Design least-privilege agents→ Implement real-time policy guards
Keynote | Threat Modeling Agentic AI Systems: Proactive Strategies for Security and Resilience
→ Design and deploy secure agentic AI systems→ Ensure safety and reliability
Will AI take over the world?
→ Develop secure AI systems→ Test AI systems for safety→ Improve AI reliability
5 essential preventative controls for Generative AI workloads | Amazon Web Services
→ Design secure and well-governed AWS environments→ Enforce consistent permissions and audit access
Miao (Mia) Zhang - Common-Sense Bias Discovery and Mitigation for Classification Tasks
→ Implement common-sense bias discovery and mitigation in image classification→ Adjust sampling weights for bias mitigation
Fire and Explosion Hazards Analysis
→ Conduct fire and explosion hazards analysis→ Estimate damages caused by explosions→ Develop prevention strategies
How Would You Implement Guardrails For An LLM Application? #Shorts #LLM #GfG #GeeksforGeeks
→ Implement guardrails for LLM applications→ Validate inputs for LLMs→ Filter outputs for LLMs
Safeguard your users and brand with W&B Weave Guardrails
→ Ensure AI system safety with guardrails→ Prevent harmful outputs from AI agents
Safeguard LLM Outputs: Test and Evaluate
→ Implement adversarial testing for LLMs→ Mitigate AI safety failures
DeepCamp AI