This challenge awaits you:

The use of virtual methods is essential for the development and validation of complex driver assistance systems (ADAS) through to autonomous driving (AD). The moving 3D simulation of a wide variety of traffic situations can replace complex vehicle tests. But creating these scenarios in virtual space can also be time-consuming and cost-intensive due to the large number of possible variations. To increase efficiency, IAV has developed a tool using an innovative AI-supported approach based on a Large Language Model (LLM) that automates scenario generation and thus significantly accelerates the process. The performance of this tool is now to be further optimized using new AI methods.

Your Tasks:

• You research the state-of-the-art regarding LLM strategies

• You select suitable strategies to optimize tool performance and/or develop new strategies

• You implement these strategies in the tool

• You test and evaluate the strategies using example and real data

• You push the boundaries of what is possible and conceivable

Necessary Skills:

• Ongoing master's degree in the MINT area

• Good knowledge of software development, e.g. with Python

• Previous knowledge in the field of artificial intelligence

Desired Skills:

• Knowledge of cloud technologies

• High level of initiative and creativity

• Enjoy trying new things