Safer, more affordable and better mobility through AI.
We build machine learning models for real mobility problems and make them available to everyone, for research and, where their licences allow it, for commercial products. From product development to vehicles and embedded systems, every model is checked for compliance, runs on your own hardware and is published with a model card that says how good it is, against what, and where it fails.
Compliant by design
Licences, privacy and the EU AI Act are checked before a model ships, not afterwards.
Runs where your data is
From a laptop down to a control unit in the vehicle. No data leaves your company, no cost per call.
Known quality, known limits
Every model card says what the model achieves, measured against what, and where it fails.
2 published, 3 in progress.
Picket (forest): CAN bus intrusion detection on a microcontroller
- Topic
- Automotive security
- Task
- can-ids
- Latest
- 0.2.0
- Usage
- commercial
Scout (large): jobs, pains and gains in mobility texts
- Topic
- Product development
- Task
- jtbd
- Latest
- 0.1.2
- Usage
- commercial
Hum (fan): machine-sound anomaly detection on a microcontroller
- Topic
- Condition monitoring
- Task
- sound-anomaly
- Latest
- none yet
- Usage
- commercial-share-alike
Pace (CNN): activity recognition from IMU windows on a microcontroller
- Topic
- Condition monitoring
- Task
- activity
- Latest
- none yet
- Usage
- commercial
Picket (MLP): int8 CAN bus intrusion detection
- Topic
- Automotive security
- Task
- can-ids
- Latest
- none yet
- Usage
- commercial
3 fields, one way of working.
Product development
Models that support product discovery in mobility, starting with jobs-to-be-done extraction.
Collection on Hugging FaceAutomotive security
Small models that detect attacks on vehicle buses and radio interfaces, built to run on microcontrollers.
Collection on Hugging FaceCondition monitoring
Small models that watch machines and movement through sensors, built to run on microcontrollers.
What the zoo stands for.
Compliance comes first
No model ships before its compliance check has passed. Sources and licences are checked and credited, personal data is kept to a minimum, and every release comes with its EU AI Act classification and a summary of its training data. If a model cannot be released lawfully, it is not released.
Built for real use cases
Every model solves one concrete mobility problem and is ready for production, with a documented interface, a stable output format and a card that says what the model is for and what it must not be used for.
Right-sized, from decision models to LLMs
Every task gets the smallest model that does the job well, from a compact decision model on a control unit to a large language model. It runs on your own hardware, so no data leaves your company, nothing depends on an outside service and there is no cost per call.
Known quality, known limits
Every model states what it achieves, at what speed on which hardware, measured against what, and where it fails. You decide whether it fits your use case before you deploy it, not after.
Open, with the licence each model needs
Models, methods, prompts and evaluation results are published under open licences, and most models may be built into commercial products. A model whose training data or components allow only non-commercial use is released under a non-commercial licence, carries the suffix -nc in its name and says so on its card and page. Each model card names the licence that applies; datasets are shared only where their own licences allow it.