We do not build apps. We build systems.
AI-enabled, cloud-native, running on your infrastructure.
The learning loop
- Partner instance
- Anonymization step
- ENSIA model lab
- Back into the system
One model, three parts: a system, an instance, a learning loop
An app is a screen. A system is everything behind it, and the models that learn from it.
AI-enabled
Every system ships with the data model and the live events a learning model needs.
Cloud-native
Kubernetes, infrastructure as code, immutable images, autoscaling, a pipeline with security gates.
Plug and play on your infrastructure
A dedicated instance: your domain, your branding, your data, your admin console. Your data never leaves your perimeter.
Daltech: live queues and appointments
Built in eighteen months by two engineers, in production on daltech.live.
- 1,656
- commits across 7 repositories
- 263
- API routes on 52 data models
- 2 to 6
- API replicas, autoscaled
- 354
- requests per second under load test, p95 67 ms, zero errors