neurocat provides innovative solutions to ensure safety compliance and comprehensive data
coverage of AI-powered perception components for autonomous vehicles. Founded in 2017,
the company bootstrapped itself through consulting on robustness in AI systems until entering
into a strategic partnership with dSPACE in 2021. neurocat has since its founding also routinely
contributed to public research and standards formulation, for instance writing the first every AI
quality standard (DIN SPEC 92001-1 / 92001-2).
neurocat’s core offering today is its aidkit software, which seeks to harmonize safety and AI for
effective AD/ADAS development. aidkit allows its users to identify vulnerabilities and data gaps
in their perception functions in order to mitigate risk across the scenes their ADAS and ADS
systems will encounter in an ODD. In addition, aidkit users can leverage generated artifacts for
model retraining and combine existent ODD attributes to discover new safe and unsafe areas.
aidkit offers also user-friendly and performance enhancing features such as easy integration
into MLOps workflow, version tracking, interim results monitoring, and horizontal and vertical
scaling to handle even larger data sets and testing within reasonable times. aidkit uses
augmented data techniques for its testing and serves as an important component-testing
element sitting alongside, and facilitating, traditional simulation and system-level testing.
