Building Trust In The AI Era
summary
Demo of Unblocked shows a multi-tenant AI developer productivity platform that integrates GitHub Confluence Slack and more and that scales to mostly enterprise customers (roughly 75–80%). Architecture decisions are concrete: per-organization encryption layered over a shared DB, Postgres row-level security for metadata, Federated SSO and session-less auth, and regular penetration testing. AI-specific controls are explicit: RAG with access checks to avoid cross-tenant leakage, prompt-exfiltration detection and classifiers, and sandboxed tool execution for web and SQL callbacks.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 0:46Developer info fragmentation
- 2:11Unblocked product demo
- 5:01Security fundamentals
- 8:36Authentication authorization
- 12:27Data security isolation
- 12:50AI security challenges
- 13:10Prompt injection risks
- 18:22Retrieval augmented generation
- 23:20Data poisoning threats
- 25:49Tool extension exploits
- 28:36AI Dev tools security
- 30:26Q&A
related talks
28mAI for Everything After Code
31mModern Architecture for AI-Augmented Development Teams
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
40mBuilding secure and efficient AI solutions in Azure with Partners
22mBuilding Secure AI Platform Infrastructure: A Zero Trust Framework
01h 11mKeynote - Securing AI: Building Trust in the Era of AI
28mSecuring AI by design
08mBuilding an AI Platform Your Engineers Actually Trust