Serve AI models at enterprise scale from endpoint to inference.
Production-grade model serving with GPU and NPU-aware scheduling, auto-scaling inference pipelines, and full observability across every request. The inference layer that connects your models to the agents, applications, and APIs that depend on them.
Turn models into production services.
Metis brings Kueue scheduling, workload execution, and serverless inference together on one platform. Launch an endpoint in one click on a ready-made inference engine, allocate resources through queues and priorities, and reclaim idle capacity. With Kubernetes-native multi-cluster operations and unified observability, you see the state of every running service in one place.
Run Metis inside the environments your enterprise controls.
The same inference platform — identical serving capability, scheduling intelligence, and observability — across your own data center and fully air-gapped environments.
On-Prem Private Cloud
Run Metis entirely within your own data center, including fully air-gapped deployments. Model inference executes on your hardware, under your network policy, with no data leaving your perimeter.
From model endpoint to high-throughput serving — a complete inference lifecycle.
Model Serving
Production model endpoint management
Deploy models as versioned inference endpoints with traffic splitting and instant rollback. Route requests across single models or multi-model configurations from a unified serving layer.
GPU + NPU-Aware Scheduling
Right compute for every inference request
Inference requests are routed to the right accelerator — GPU or NPU — based on workload profile, hardware availability, and cost policy. No manual affinity rules. No hardware misconfiguration.
Auto-Scaling Inference Pipelines
Capacity that follows demand
Scale inference capacity automatically with real workload demand — up during peak periods, down during idle windows — without provisioned waste or manual capacity planning.
Full Inference Observability
Complete visibility across every endpoint
Real-time metrics on latency, throughput, token usage, hardware utilization, and error rates — surfaced per model, per endpoint, and per request. Identify bottlenecks before they become incidents.
How Metis fits into the stack.
Metis is the inference layer that models built in Maxis deploy into.
Metis
Production inference platform providing model serving, endpoint management, GPU/NPU-aware scheduling, auto-scaling, and full observability. The layer where trained models become production services.
When to use Metis.
When several teams need to share an in-house model through an API
Route requests through a single serving layer.
When traffic changes by the hour and capacity is hard to fix in advance
Scale down when idle and start up automatically when requests arrive.
When you need to change models while the service keeps running
Versioning and traffic splitting help switch models with less service impact.
When inference requests must not leave the company
Endpoints stay inside your own environment.