AutonaAUTONA
Division 03 · AI

Intelligence engineered for the physical world

Autona AI develops foundation models for perception, planning and control that run at the edge — enabling robots, aircraft and defense systems to make sub-millisecond decisions without cloud dependency.

Capabilities

Perception Models

Multi-modal transformers fusing camera, LiDAR, radar and RF signals.

World Models

Predictive models of scene dynamics for planning under uncertainty.

Autonomous Planning

Hierarchical task planners with formal safety envelopes.

Edge Deployment

Quantized runtimes for Jetson, Hailo and neuromorphic silicon.

Continuous Learning

Federated learning across fleets with differential privacy.

Assured AI

Model provenance, red-team testing and MIL-STD-882E safety cases.

Platforms

Deployed systems

Foundation Model
Autona.Cortex

40B-parameter multi-modal model tuned for autonomous systems and geospatial reasoning.

Runtime
Autona.Edge

Deterministic inference runtime with hard real-time scheduling guarantees.

MLOps
Autona.Studio

End-to-end lifecycle management with automated evals and drift detection.

Reference specifications
Model families
Cortex-Vision · Cortex-Plan · Cortex-Control · Cortex-Language
Latency target
< 8 ms perception loop on Orin AGX
Assurance
ISO/PAS 21448 SOTIF · EU AI Act high-risk conformance
Next step

Discuss a deployment

Contact division