NLP/LLM& AI Agent科学家_CR
你的任务
We are seeking an applied Research Scientist to join our team.
In this role, you will lead the design and development of large language models (LLMs)-based intelligent agent systems tailored for Bosch's industrial manufacturing scenarios.
You will leverage LLMs, agent development frameworks, and advanced NLP techniques to enable high-precision task automation, intelligent reasoning, and decision support.
You will architect and implement end-to-end agent capabilities—including function calling and tool use, memory mechanisms, task planning, and retrieval-augmented generation—to combine general AI capabilities with Bosch's deep industrial knowledge.
A key part of this role is driving post-training and fine-tuning efforts (e.g., instruction tuning, reward modeling, domain adaptation) that transform foundation models into high-accuracy, production-ready agents for real-world factory use cases.
Beyond solution development, you will continuously scout and evaluate the latest advancements in LLMs, agent frameworks, and knowledge-enhanced AI to ensure our solutions remain state-of-the-art.
By integrating cutting-edge research with Bosch's domain expertise, you will help close the "last mile" of AI deployment in industrial environments—delivering robust, reliable, and high-precision agent performance in production-level applications.
您的个人资料 • Design, develop, and optimize LLM-powered agent systems—including memory, planning, and reasoning capabilities—for high-precision task automation and decision support in industrial manufacturing scenarios.
• Execute post-training and alignment of LLMs—including supervised fine-tuning (instruction tuning, domain/task adaptation), parameter-efficient tuning methods (e.g., LoRA/QLoRA), reward modeling, and preference optimization—to deliver controllable, high-accuracy behavior in industrial applications.
• Advance RAG systems beyond baseline implementations by optimizing retrieval, chunking, reranking, and grounding to achieve high-accuracy, domain-adapted performance in industrial scenarios.
• Collaborate with domain experts to transform industrial data into actionable insights using advanced NLP techniques
• Validate and benchmark agent performance in production-like environments, ensuring robustness and efficiency
• Contribute to scientific publications, patents, or technical whitepapers as part of Bosch's innovation initiatives
Basic Qualifications
• PhD or Master's degree in Computer Science, Artificial Intelligence, NLP, or related fields, with 3+ years of working experience in building real-world NLP or agent systems
• Proficient in Python and widely used LLM toolchains—covering model development (Transformers, PyTorch) and agent orchestration frameworks (LangChain, LangGraph, or equivalent).
• Solid understanding of LLM post-training and alignment workflows, including supervised fine-tuning (e.g., instruction tuning, domain adaptation) and reward/preference model optimization, with hands-on experience in at least part of this pipeline.
• Solid understanding of agent architectures (e.g., RAG, memory, tool use, planning) and their application in high-precision environments
• Proficiency in English for technical communication and collaboration in interdisciplinary teams
• Understanding of the deployment challenges of LLM/AI systems in production environments
________________________________________
Preferred Qualifications
• Experience applying LLM agents to industrial manufacturing domains
• Familiarity with knowledge graph related technologies- a plus
• Demonstrated contributions to top-tier AI/ML/NLP research (e.g., ACL, NeurIPS, ICML, ICLR)
• Some exposure to CV or multimodal models is a plus
• Design, develop, and optimize LLM-powered agent systems—including memory, planning, and reasoning capabilities—for high-precision task automation and decision support in industrial manufacturing scenarios.
• Execute post-training and alignment of LLMs—including supervised fine-tuning (instruction tuning, domain/task adaptation), parameter-efficient tuning methods (e.g., LoRA/QLoRA), reward modeling, and preference optimization—to deliver controllable, high-accuracy behavior in industrial applications.
• Advance RAG systems beyond baseline implementations by optimizing retrieval, chunking, reranking, and grounding to achieve high-accuracy, domain-adapted performance in industrial scenarios.
• Collaborate with domain experts to transform industrial data into actionable insights using advanced NLP techniques
• Validate and benchmark agent performance in production-like environments, ensuring robustness and efficiency
• Contribute to scientific publications, patents, or technical whitepapers as part of Bosch's innovation initiatives
Basic Qualifications
• PhD or Master's degree in Computer Science, Artificial Intelligence, NLP, or related fields, with 3+ years of working experience in building real-world NLP or agent systems
• Proficient in Python and widely used LLM toolchains—covering model development (Transformers, PyTorch) and agent orchestration frameworks (LangChain, LangGraph, or equivalent).
• Solid understanding of LLM post-training and alignment workflows, including supervised fine-tuning (e.g., instruction tuning, domain adaptation) and reward/preference model optimization, with hands-on experience in at least part of this pipeline.
• Solid understanding of agent architectures (e.g., RAG, memory, tool use, planning) and their application in high-precision environments
• Proficiency in English for technical communication and collaboration in interdisciplinary teams
• Understanding of the deployment challenges of LLM/AI systems in production environments
________________________________________
Preferred Qualifications
• Experience applying LLM agents to industrial manufacturing domains
• Familiarity with knowledge graph related technologies- a plus
• Demonstrated contributions to top-tier AI/ML/NLP research (e.g., ACL, NeurIPS, ICML, ICLR)
• Some exposure to CV or multimodal models is a plus
