with the help of the NanoThink.AI dream team
About us and our mission
We are able to solve complex and creative tasks of developing software and hardware for devices of various purposes, especially where AI is executed on the device itself (microcontroller, EdgeAI, TinyML technologies).
Our mission is to assist companies which are looking to obtain a prototype / first version of a device to confirm the technical feasibility of the idea, especially for devices which operate using artificial intelligence (neural networks).
What we offer...
Together with the customer, we go from the initial idea to creating a prototype or first experimental batch:
- we check the technical feasibility of the idea;
- we create a mock-up, conduct tests;
- we develop a technical specification;
- we build a prototype and draft all the documentation for creating a small batch of devices.
We can prepare for industrial or large-scale mass production, and we also know where we can carry out industrial device design with our strong industrial partnerships in China.
We work hard every day to make our clients' lives clients better and happier
- Start-ups
Who have a product idea, they know what the customer requires, they want to build a prototype quickly to test the idea or in order to obtain funding
- Manufacturers of electronic devices
Who want to integrate artificial intelligence into their products (to make or modify their smart vacuum cleaner, refrigerator, etc.)
- Custom electronics development companies or R&D departments
Who develop devices for their clients/companies, but who lack resources or experience in the field of TinyML
We are interested in clients involved in the following industries
- Consumer electronics and IoT devices: for smart home, personal assistants, robotic vacuum cleaners, fitness trackers and other products that can improve a person's daily life with the help of artificial intelligence.
- Industry 4.0: production process automation, unmanned vehicle solutions, forecasting and resource management systems.
- Medical technologies and healthcare: development of smart medical devices for monitoring and controlling patient health, creation of decision support systems in the process of patient treatment and rehabilitation.
- Security: video cameras with biometric recognition, access systems, smart locks.
- Robotics: for processing sensory data on board robots.
- Measurements and control: smart sensors and meters, measurement systems, predictive analysis.
And so on, and so forth: there are no restrictions. Any electronic device, or any device containing an electronic unit, and ideally also with the task of implementing AI.
The creation of smart devices requires competencies in several technological areas at the same time: embedded, ai, web/mobile app, electronic circuit board design, equipment prototyping by combining several boards/sensors/technologies.
As specialists in Edge AI / TinyML, we have all these competencies, and we do not create individual components, but rather a complete solution for the client's task.
Also, our cooperation with various manufacturers of processors and accelerators gives us another advantage - we have practical experience with a wide range of hardware platforms, and can choose the optimal platform for the client (or quickly adjust to the platform already selected by the client).
Typical project
plan
• Task analysis
• Architectural design
• Electrical design:
◦ ◦ Component base selection
◦ ◦ Circuit schematic creation
◦ ◦ PCB layout development
◦ ◦ PCB ordering
◦ ◦ Assembly and testing at the hardware level
• Mechanical design
• Software development
• Comprehensive testing: hardware and software.
Device for classification of anomalies based on the reservoir computing algorithm for ESP32-C3
The module identifies and classifies anomalies in the operation of equipment with rotating elements, reducing the cost of creating predictive analysis systems.
The low cost of the module ($4.20), along with its low power consumption, does not require the cost of model development and training (including markup and preparation of the data set): it is trained "on the fly" directly on the equipment. Re-computing, Echo State Network, On-Device Learning (ODL), tinyxml, PCB Design and Layout, ESP32, OTA updates
Recognition of objects in a video stream on an autonomous device (on a microcontroller) without connecting to a unit for controlling the monitoring of wild animals
A device that allows you to configure certain alignment parameters for each output
Software for devices
- BSP, drivers
- Firmware
- Applications (including mobile and cloud-applications)
- Execution infrastructure deployment
- OTA updates infrastructure design
- Neural network and hardware platform selection to solve your task with the best performance/cost ratio
- Remote Continuous Integration on the real hardware set-up and support
PCB-design
- Printed circuit board design
- Hardware prototyping
- Ethernet, USB, FPGA, MCU, DRAM, LCD
- Eagle (Autodesk Fusion 360), KiCAD, Altium Designer, OrCAD
Development tools
- Compilers, debuggers, SDK, IDE, toolchain development
- For AI chipmakers - TVM and LLVM adaptation for specific chip for the best AI tasks performance
- Stack: LLVM, TVM, TensorFlow, Eclipse, SMP, OpenMP, POSIX
- Platforms: ARM, PowerPC, MIPS, Custom platforms
We want our clients to also be our friends, so we are happy to answer your questions.
Email: hello@nanothink.ai
United Kingdom, London, EC1N 8LE, Hatton Garden 63/66
