Akamas

Akamas began with a recognition drawn from more than 20 years of performance engineering at Moviri: modern applications had become too interconnected and configuration-heavy for engineers to discover optimal settings manually. As Kubernetes, cloud infrastructure, and AI inference expanded the number of interacting parameters, conventional monitoring could reveal problems but could not consistently determine which changes would improve performance, reliability, and cost without introducing risk. The founders therefore created Akamas as an autonomous, continuous optimization company built around patented AI and practical performance-engineering expertise.

Today, Akamas operates from Milan and serves developers, DevOps teams, platform engineers, SREs, and performance engineers. Its products include Akamas Vivo for full-stack Kubernetes optimization, Akamas Studio for testing and tuning arbitrary technology stacks, and Akamas Inference Optimizer for improving GPU utilization and token throughput in self-hosted LLM inference. The platform connects with observability systems such as Dynatrace, Datadog, and Prometheus, analyzes application, cluster, and infrastructure data, and produces evidence-backed configurations that teams can review, approve, and apply through GitOps. Akamas extends this application-aware approach to JVM, Node.js, .NET, vLLM, GPU, autoscaling, and node-group settings, helping enterprises reduce manual tuning while improving service performance and infrastructure efficiency.