Building elastic, AI-assisted infrastructure — from dynamic Kubernetes autoscaling to intelligent developer tooling — in partnership with leading innovators.
Partnerships
Deep technical collaborations with companies pushing the boundaries of cloud-native computing and AI-driven development.
More requests, no limits.
Dynamic autoscaling that honours your requests and removes artificial limits.
Collaborating on the Express Compute Platform — a next-generation Kubernetes-based infrastructure delivering true dynamic autoscaling for both CPU and GPU workloads. The platform targets demanding CI/CD pipelines as well as training and pruning of SLMs and LLMs, with transparent workload migration and resource protection built in. MemVerge worked together with NVIDIA on CRIU for GPUs, bringing checkpoint/restore capabilities to GPU workloads and enabling seamless live migration of GPU processes.
A multi-project partnership addressing both Kubernetes operations and Java developer experience, with AI woven throughout the stack.
Gives administrators and end-users auditable, controlled access to ephemeral containers on Kubernetes. Enhanced AI analysis processes captured results from profilers and bin-packer data collectors, delivering actionable insights that close the loop between observation and optimisation.
github.com/codriverlabs/KubeCoDriver-ce →AI-enhanced compression and transfer of Java artifacts, dramatically optimising the Java developer experience during development. Machine-learning models guide streaming strategies to reduce latency and bandwidth overhead.
jarz-streaming.net →Open Source — AWS Ecosystem
Internally driven open source projects that extend the AWS ecosystem — built for the community, released under open licenses.
EKS Distro fully operational on a single EC2 instance, spinning up in under 2 minutes — with native Karpenter support for node provisioning and a compatibility layer for EKS Pod Identities, making it a drop-in development and CI target for production EKS workloads. Designed to dramatically lower the cost and complexity of running EKS-compatible environments outside of managed clusters.
A Maven resolver extension that mounts S3 buckets as local artifact repositories using Mountpoint for Amazon S3 — enabling direct, low-latency access to Maven artifacts stored in S3 without a proxy or Nexus-style mirror.
github.com/plasticity-of-cloud/ecp-maven-s3-split-resolver →Helper tooling that brings Amazon DCV (formerly NICE DCV) remote desktop support to Ubuntu 25.04 and newer, where official packages are not yet available. Automates dependency resolution and compatibility shims.
github.com/plasticity-of-cloud/aws-helper-tools →GraalVM native-image optimisations powered by AI recommendations with a closed-loop Profile Guided Optimisation (PGO) pipeline. The SDK continuously collects runtime profiles, feeds them into AI-driven analysis, and applies the resulting compilation hints — tightening the feedback loop between production behaviour and build-time decisions.