Handling Sensitive Data with Cloud and Local AI

Course details

In this comprehensive course, full-stack software developer, speaker, and Python enthusiast Ronnie Shear teaches you how to securely integrate AI solutions with sensitive data. Learn to evaluate trade-offs between cloud-based and on-premise inference platforms, ensuring your decisions align with strategic goals. Explore ways to implement data anonymization and security best practices to protect user privacy and comply with regulations. Discover strategies for building a robust AI safety framework to mitigate risks related to privacy, fairness, and misuse. Ideal for architects, decision-makers, and developers, this course empowers you to make informed choices about AI models and tools, employ an AI safety framework, enhance trust in your organization's AI systems, and safeguard your data assets.

Learning objectives
Evaluate trade-offs between cloud-hosted and on-premise AI solutions.
Implement frameworks for safer AI use.
Apply anonymization and security best practices.
Evaluate deployment options for sensitive data.
Select appropriate AI models and tools.

Concepts

Data Sovereighty and AI

Privacy controls in popular AI assistants
Understanding AI and data safety
Build a safety framework for responsible AI use
Choosing an inference platform
Visualizing LLM risks - Create an interactive UI
Build it - Implementing the dual LLM pattern

Assistants as a Service

AI assistants - Secure options for sensitive data
Configure AI assistants for maximum data security
Apply data anonymization for safe AI interactions
Implement an LLM proxy with liteLLM

Cloud-Hosted Options

Cloud AI - AWS Bedrock and Azure Foundry
Run an LLM using cloud inference

Local AI (On-Premises)

Considerations of self-hosting
Deploying an on-premise assistant
Vibe code securely on a NVIDIA DGX

Conclusion

Next steps