Ainomaly, a deeptech company, makes software safe with artificial intelligence and machine learning to reduce the risk of autonomous driving. Spot on there, since the risk of not being able to guarantee that safety today, is a bigger issue than delivering the autonomous cars themselves.
Meeting Karl Meinke and David Berglund makes me realize that the play-date they had in Karl’s kitchen when they came up with the idea after 10 years of research at the KTH Department of Computer Science is not an accident. A 30-minute talk probably symbolizes these guys with ingredients like safety, risk, proximity, credibility and a bunch of great laughs.
But also these 30 minutes reveal a fantastic will to make a difference to their customers and bring future mobility closer, quicker and safer.
The challenge in developing safe autonomous vehicles lies in the fact that the unlikely will happen. The development of autonomous vehicles is today driven by demands for sustainable solutions, the consumer’s need for better transport and society’s development towards environmentally friendly and smart cities. A development that in itself creates new challenges to solve.
Autonomous cars by autonomous tests
In order for systems to be reliable, they must have been tested even in the most unexpected of scenarios. Software is constantly updated and it’s impossible to carry out complete tests for each new change if most of these tests do not become completely virtual and autonomous.
The solution is to simulate car environments in a virtual system that guarantees that you will find the demanding edge case, i.e. the many but important exceptions that almost never happen, but which will still eventually happen sometime.
Don’t let the bugs (anomalies) bite. A name and a fact
Ainomaly apply active machine learning based on a branch of AI called eXplainable AI (XAI). XAI is a domain where creation of datasets and analyses can be self-certifying. Ainomaly has created an ethical application of AI that enables customers to build safer software systems by identifying unexpected behavior, anomalies and critical errors more quickly.
Coopetition – rather than competition. What does that mean?
We are not here to reinvent the wheel. We provide an engine for test cases, i.e. a test content generator (for a simulation environment). We plug that engine into the customer’s test environment. Hence the competitors are also collaborators, which brings an extra value - coopetition.
Being an emerging company working with the global companies, what’s your conclusion on how to find your way into the organisations?
Being part of MobilityXlab is of course the main factor for us getting the contacts we need in their partner organisations. But it’s also important to decide whom really is your coworking department. For us it’s absolutely to be working closely with the research and the production departments in the organisation.
If you would give your best tips to other startups in your position...
It’s very important to reach credibility with your customer and in order to do that you really need to understand their problem on many levels. And then we mean their real problem. Don’t grab the first thing you hear.
If reaching credibility is important, what would you ask for from the big companies?
That would definitely be to ask them to let us in and to work close with us emerging companies. It’s only by doing that, working with the production teams, when we together get the best possibilities for ever finding the real problems, e.g. the bugs. Before they find you.
We want to work broadly, contributing to their solutions on several levels within the global companies. It might be a hard way to do business, but once you’re on the right spots, you’ll reach credibility, which is one of the key success factors for an emerging company like Ainomaly.
The feeling of contributing to something bigger
Ainomaly contribute to two or more of the global sustainability goals by building data tools for sustainable and ethical applications, with high business value. ”Climate action ” and ”Reducing the inequalities” are targeted goals that they intend to contribute to.
Headquarter: Stockholm, Sweden
Business: Making software safe with explainable AI to reduce the risks of autonomous driving.
Co-Founders: Karl Meinke (left on picture), David Berglund (right on picture)
Follow our startup stories on mobilityxlab.com or LinkedIn!
MobilityXlab is an unconventional collaboration platform where large and emerging companies with
pioneering ideas drive innovation together. Our founders are CEVT, Ericsson, Veoneer, Volvo Cars, Volvo Group, Zenuity and Lindholmen Science Park.
Text och foto: Christina Thulin