DevOps for Scientific Software - Tools, Practices, and Automation Strategies

This video is posted in 2026-04-03.

Scientific software development often struggles with reproducibility, complex dependencies, and scaling across HPC and cloud environments. This session shows how modern DevOps practices can overcome these challenges and accelerate research. We’ll explore how GitHub Actions automates testing and deployment, while Docker ensures consistent, reproducible environments across teams. For large-scale workloads, Kubernetes provides robust orchestration, and Karpenter delivers intelligent autoscaling to optimize cost and performance dynamically. We’ll also cover observability and monitoring with Prometheus and Grafana, giving teams real-time insights into resource utilization and system health. By combining CI/CD, containerization, orchestration, and monitoring, attendees will learn practical strategies to make scientific software faster, more reliable, and ready to scale efficiently in HPC and cloud environments.