Keynote: MLOps on Highly Sensitive Data - Strict Confinement, Confidential Computing, and Tokenization Protecting Privacy

The presentation discusses the use of secure MLOps in the life science industry, with a focus on protecting patient privacy and complying with industry standards.

• Tokenization is used to protect patient privacy by changing personally identifiable information to a token based on a hardware security key.
• Strict confinement features of micro-kubernetes distribution are used to ensure tamper-proof tokenization.
• Confidential computing is used to expand local Kubernetes clusters in a safe way by creating a VM on a public cloud and utilizing open enclave and open source projects to configure the confidential compute and underlying hardware features.
• The benefits of using public clouds for research use cases are discussed, including the ability to spike up capacity when training a bigger model.
• The presentation emphasizes the importance of using secure MLOps to comply with industry standards and protect patient privacy.

The presenter discusses the use of tokenization to protect patient privacy, explaining that it is different from encryption in that it looks inside the data set and changes personally identifiable information to a token based on a hardware security key. They illustrate this with a fictitious patient data set, showing that the biological information is still usable but the name was tokenized. They also emphasize the importance of using secure MLOps to comply with industry standards and protect patient privacy.

MLOps is used in various organizations, that operate on very sensitive datasets. Pharmaceutical and life science companies handling human DNA samples, healthcare institutions training models on patient data, or highly regulated environments like telecom and financial companies. Many users are afraid that cloud-native would expose them more to vulnerabilities, data leaks, or other security issues. In reality, it's just the opposite. With Kubernetes and its ecosystem - Kubeflow, strict confinement for K8s using AppArmor profiles, confidential computing in case you run your workloads on the public cloud and blockchain-based tokenization you can achieve very safe and compliant setup. On the talk you will see a case study of a LifeSciences company creating customized treatments based on DNA, utilizing above mentioned technologies to run complex hybrid/multi-cloud MLOps using Kubernetes and Kubeflow.