Accelerating privacy & security in AI with Amazon Bedrock and Tines
summary
Tines builds an end-to-end solution that runs language models inside a VPC using Amazon Bedrock over PrivateLink so no data leaves the infrastructure and Bedrock enforces no-training/no-logging guarantees. A first shipped feature, Automatic mode, uses Claude 3 Sonnet (then Haiku for speed) to generate transformation code; the follow-up AI action embeds LLM steps across workflows (example: near-perfect CEO‑fraud scoring on suspicious emails) and removes the security/privacy vs. accuracy tradeoff. Bedrock+PrivateLink is presented as the production-ready path.
AI-written summary · about
chapters (11)
- 0:00Defining important workflows
- 1:01Powering automation platforms
- 2:24The intersection of AI and automation
- 3:34Challenges in building AI features
- 5:40Privacy and security trade-offs
- 7:37The ideal infrastructure solution
- 8:37Building and prototyping for trust
- 10:39Leveraging Amazon Bedrock
- 12:47Strategic benefits of secure AI
- 13:55Implementing the AI action
- 15:14Conclusion and final advice
related talks
55mSecurely build generative AI apps & control data with Amazon Bedrock
24mBuilding GenAI applications with Amazon Bedrock
20mData is Your Differentiator: Building Secure and Tailored AI Systems
31mLamRAG: From Data to Constructive Insights using Amazon Bedrock
22mObservability in AI-Powered Applications with Amazon Bedrock
29mCustomizing genAI applications for your business
20mCollaborate across the cloud, Amazon Bedrock, gen AI, AWS DeepRacer
38mGenerative AI - Architectures and applications in depth