AI / ML Engineer

• \n University degree in Computer Science, Engineering, or a related field.\n \n
• \n Several years of experience in applied machine learning and model engineering.\n \n
• \n Proficient in PyTorch and Transformers frameworks.\n \n
• \n Expertise in model compression, alignment, and fine-tuning of small GenAI and ML models.\n \n
• \n Skilled in building evaluation suites and performing statistical analysis of model behaviour.\n \n
• \n Practical knowledge of edge device constraints such as RAM, bandwidth, cache, and cold start.\n \n
• \n Professional English proficiency, German is a plus.\n \n

\n

\n Are you ready to take proactive action with a hands-on mentality, strong problem-solving skills, and excellent cross-functional communication?\n \n

n Artificial Intelligence and machine learning are at the cutting edge when it comes to shaping next-gen mobility in areas like automated driving. But it takes real intellectual leadership and expertise to identify and then harness the AI trends and technologies that will transform the way people travel for good.\n \n

\n We are an international team of experts shaping the future of intelligent mobility by developing and refining edge GenAI models for the BMW Intelligent Personal Assistant. Together, we deliver customer impact through innovative model engineering and robust data pipelines.\n \n

What awaits you?

\n \n

• \n You select and adapt base models for voice assistant and intelligence tasks.\n \n
• \n Furthermore, you produce clean, reproducible training, finetuning and evaluation pipelines to refine models for the problem at hand.\n \n
• \n You implement required compression strategies such as quantization, pruning and distillation to enable cost-efficient inference.\n \n
• \n Additionally, you collaborate with performance engineers on hardware-aware model variants and performance.\n \n
• \n You co-own latency and memory KPIs and provide models that meet budget requirements on the target hardware.\n \n
• \n You document design choices, trade-offs and model cards for audits and compliance.\n \n

\n