Experience hybrid cloud AI/ML at scale

Data Science practices face challenges of scale, reproducibility, on-premise and cloud data silos, and inconsistent deployments across environments. Red Hat has played a key role in solving these issues for agile software development, especially with DevSecOps principles and practice. Now, Red Hat is applying that experience and perspective to principled Machine Learning Model Operationalization Management practices at scale. Please join us as we use Red Hat OpenShift Container Platform and ML for the detection of pneumonia from chest x-rays. This cloud-agnostic solution uses popular ML tools for running large and distributed workloads across disparate multicloud data sets. Compute resources are spun up on-demand, and scale to 0 as compute jobs reach completion, while edge-deployed ML models that show low degrees of certainty are retrained in centralized data hubs and continuously deployed back to the healthcare environment. At scale, this data pipeline reduces compute latency, and facilitates continuous improvement and deployment for time-sensitive imaging and diagnostics. Red Hat has a proven and global track record of helping organizations operationalize AI/ML solutions in multicloud and hybrid cloud environments using open source tools for data harmonization, distributed AI and ML workflows, and the emerging practice of MLOps.