Offline training software
Label images, train and retrain vision models on local compute within your own infrastructure.
Build, improve and deploy mission-specific AI on edge devices. Designed for the field. Even when the cloud is out of reach.



Jump Edge offers a sovereign Edge AI framework for defence and security, with vision models supplied and trained as part of the solution.
We train, adapt and optimize vision models for your target hardware—reducing model size and balancing recognition quality, inference speed, memory and power requirements. Models, runtime and TAK integration are aligned as one solution.
The framework connects the full lifecycle, from model training and validation to local inference and improvement with field data, within customer-controlled infrastructure.
Label images, train and retrain vision models on local compute within your own infrastructure.
Base models for visual and thermal recognition, trained and optimized for your use case and device.
A tactical app for local vision AI, with TAK integration and a TAK plugin to connect validated observations.
Prepare and transfer validated models to compatible edge devices through an agreed local deployment path.
Jump Edge builds on Jump’s software delivery experience with:



Discuss your Edge AI use case with us. In a 30-minute introductory conversation, we explore your situation and what a sensible next step could be.
What needs to work, where, and for whom?
Available data, devices, connectivity and infrastructure.
What to validate and whether a pilot makes sense.
Jan-Maarten VerweijCo-founder · JumpJust getting to know us? Meet Jump