During the development of ADAS it is essential to identify certain scenarios in the available datasets. For this purpose vision-language transformer models, like CLIP, could be utilized using open source visual ADAS datasets. The task includes the set up of a state-of-the-art vector database (e.g. Milvus) to store and query the extracted features based on which the scenarios should be defined. Required skills: - Python - Docker - (optional) Experience with LLM fine-tuning

During the development of ADAS it is essential to identify certain scenarios in the available datasets. For this purpose vision-language transformer models, like CLIP, could be utilized using open source visual ADAS datasets. The task includes the set up of a state-of-the-art vector database (e.g. Milvus) to store and query the extracted features based on which the scenarios should be defined. Required skills: - Python - Docker - (optional) Experience with LLM fine-tuning