VECTORIO“Your cars record terabytes; your pipeline starts after the upload”
Test-drive logs, filtered at the garage.
KPIs and scenario filters computed before data is uploaded
- Test cars fill disks faster than we can upload, so we ship drives.
- We cannot find the 10 relevant seconds in 10 petabytes.
- Our supplier data and test logs may not leave Germany.
- KPIs and scenario filters computed before data is uploaded
- Anonymised footage before storage
- Release-over-release KPI comparisons that are reproducible
From today to the first result.
shipping drives from the test cars
Test cars fill disks faster than the upload runs, so drives travel by post, and the relevant seconds are hard to find in the pile.
KPIs and scenario filters are computed at the test bench before upload, and footage is blurred at 18.7 ms per frame at 30 fps with 0 % drops before it is stored.
Compute one KPI at the garage on the log of one test car.
one test carLive camera, on-device detection. Step by step.
A live camera stream is analysed by a local model; only confident detections are kept and sealed as evidence.
- CameraRTSP Stream Input
- Detect · MonitorONNX Classifier · Video Viewer
- Row FilterETL.Filter
- EvidenceEvidence Collector
- PackageEvidence Pack
camera-input → Video.RTSP.Subscriber
object-detect → Classification.Onnx
confidence-filter → ETL.Filter
evidence-collect → Forensic.EvidenceCollector
evidence-pack → Forensic.EvidencePack
viewer-output → Video.Viewer- CameraReads a live camera stream.
- DetectRuns an ONNX model on each row or image.
- MonitorShows frames live in a window.
- Row FilterKeeps only rows that satisfy a condition.
- EvidenceHashes rows and batches into a tamper-evident evidence manifest.
- PackageSeals a manifest into a verifiable evidence package.
A real workflow from our end-to-end test suite; the designer view is illustrated.
What you get. From the first pilot.
KPIs and scenario filters computed before data is uploaded
Anonymised footage before storage
Release-over-release KPI comparisons that are reproducible
Why your tools fall short.
Spark/Databricks costly at petabyte scale and cloud-bound
ROS bag tools work one file at a time
Hyperscaler automotive stacks conflict with IP and sovereignty needs
- ADAS/AD validation volume
- Privacy rules for camera data
- Cost of cloud storage and egress
- Supplier/OEM IP protection
Your pilot. Three steps.
Your data shape
Tell us what data you have, where it lives and what the result must be.
A scoped pilot
On your data and your hardware, with the success measure written down first.
A clear decision
You see the result against that measure — and keep every workflow and record.
Related roles in Research & geodata.
Research instruments & big-science data
Instrument data. Answers in minutes.
Climate, weather, geodata & earth observation
Satellite scenes to a signed city map.
Digital pathology & microscopy imaging
Slides classified. Results signed. On-prem.
Media & streaming analytics
Real-time media, no dropped frames.
Bring us your hardest data problem. Let's solve it together.
A short pilot on your data and your hardware. We agree on what success looks like before we start.