Unlocking Cross-Organizational Insights - Practical MPC for Cloud-Based Data Analytics

Explore a practical conference talk that demonstrates how organizations can implement Secure Multi-Party Computation (MPC) systems for collaborative data analytics while maintaining strict privacy standards. Learn about the development of user-friendly features that enable developers without cryptographic expertise to perform private analytics using intuitive Python-like code, built on open-source frameworks including Carbyne Stack and MP-SPDZ. Discover key modifications made to existing frameworks to achieve stable deployment, including the implementation of semi-honest security models that offer more practical and cost-effective solutions than malicious security models in real-world scenarios. Examine practical considerations around cost optimization and performance enhancement, including infrastructure deployment strategies and algorithm-level improvements that enable complex analytics. Understand how these MPC solutions can be applied in practical contexts such as the AdTech industry, with insights from lessons learned during real-world implementation. Gain knowledge about making secure cross-organizational data analytics accessible to developers without specialized MPC expertise, presented by industry experts who have successfully deployed these systems in production environments.