Startup Profile
PanocularAI

Introduce yourselves

Tell us in three sentences who you are

Our mission is: We make embedded AI efficient and reliable with ease!
We are an early-stage startup that brings the AI revolution directly to edge-devices without compromising performance, privacy or efficiency. Our ground-breaking technology lays the foundation for next-generation intelligent systems that are efficient, secure and reliable.

How did the idea come about and what challenges had to be overcome?

The idea behind PanocularAI emerged from years of research at TU Darmstadt. Surprisingly, the idea for PanocularAI originated in two minds that were not connected until a few years ago. Both founders realised in their fields that an AI that is easy to adapt and roll out, versatile in its connection options and yet efficient in its execution does not yet exist. The path of two similar ideas into a common vision and the corresponding applications, building a strong team and the financial security in this new domain have been the three biggest construction sites we have had to struggle with so far, now comes the technical work.

How does PanocularAI work and how does AI support this project?

We are working on fundamental technologies to enable the execution of high-quality AI on edge-devices, so the question should rather be, which project can we support with our AI… Here the answer is quite easy: Any user/application that needs an edge-based AI solution will be able to benefit from our technology.

How will the technical implementation of the AI software be driven forward and what are the next milestones in development?

We have two milestones to make our technology presentable:

Smart-Cam:
Our framework will be used to create a multimodal AI that will be evaluated in the context of a smart camera application.

Drone:
Our framework will be used for the creation of a multimodal and multitask AI, which will be deeply integrated into an existing system.

A third milestone deals with the integration of our framework into existing tool environments for the creation of accelerated AI solutions.

How does your technology differ from other competitors in the field of autonomous vehicle perception with regard to the use of drones, among other things?

We have an efficient multimodal AI that is also capable of multitasking in the context of the use case. With a special optimisation pipeline that optimises the AI model and generates efficient code for the target hardware, we can make environment perception more energy-efficient and safer and, in the case of drones, enable not only a better assessment but also a longer flight time.

Let’s take a look into the future: Where could the project be in five years’ time?

In five years, we will see our framework and the solution from it in various AI architectures and, above all, we will have enabled the world of mobile robotics to develop efficient, safe and cost-effective solutions for the problems of the coming decades.