Technologies

Tools we build with

Grouped by the layer of the system they belong to. Names here are the stack, not a certificate or a customer list.

Connected stack

AIModelsModelsDevicesDevicesSoftwareSoftwareDataData

Devices, software, data, and AI models meet in one processor.

01

Embedded

Devices and firmware we use when the software has to live on the hardware.

STM32

ARM Cortex microcontrollers for tighter timing and industrial I/O.

Raspberry Pi

Linux boards for gateways, vision hosts, and ROS 2 experiments.

C/C++

Firmware and other device code where memory and timing are part of the design.

02

AI / ML

Languages and libraries for models, data preparation, and applied automation.

PyTorch

Model development and training when the team needs to see and change the network.

TensorFlow

Training and export when a deployment target expects that stack.

scikit-learn

Classical models for tabular sensor and operations data.

03

Computer Vision

Capture, classical processing, and detectors for a defined camera task.

YOLO

Real-time detection and counting models for known object classes.

04

Robotics

Robot software and simulation used before a platform is asked to move on a real floor.

Gazebo

Simulation of a route, a sensor layout, and a blocked path.

05

Web

Interfaces and APIs for operators, this website, and services that sit above the devices.

Next.js

This website, and a base for server-rendered product interfaces.

FastAPI

Python APIs between models, devices, and web applications.

06

DevOps

The everyday tools for building, running, and keeping history of the software.

Git

Source history for firmware, models, and application code.

Linux

The host for gateways, build machines, and servers, including this Ubuntu setup.