Jan Krepl developed secure private-network access for a machine learning service by integrating the ML ECS workload with a private Application Load Balancer in the openbraininstitute/aws-terraform-deployment repository. He updated ingress rules and Terraform modules, ensuring that ML traffic was routed exclusively through the private ALB, which enhanced network isolation and reduced exposure to public networks. Using AWS, Terraform, and HCL, Jan provisioned and wired the necessary infrastructure as code, associating security groups to enforce internal-only access. This work improved operational reliability and security for ML workloads, demonstrating a focused application of DevOps and networking principles within a short project timeframe.