Data Scientist (m/f/d)

Advantest; Deutschland; Vor Ort; EUR 80.000 - 100.000; Vollzeit

Jobbeschreibung

• As a Data Scientist (m/f/d), you will collaborate with senior team members to develop and deploy machine learning solutions, with a focus on Large Language Models (LLMs), for semiconductor testing and electrotechnical systems.
• Assist in designing and implementing LLM-driven tools to automate data analysis, technical documentation processing, and defect prediction workflows for the 93K IC Test platform.
• Work under guidance to preprocess datasets, fine-tune open-source LLMs (e.g., LLaMA, Mistral), and integrate retrieval-augmented generation (RAG) systems into testing pipelines.
• Contribute to MLOps workflows for model training/evaluation using Python frameworks (PyTorch, Hugging Face) and cloud platforms (AWS/Azure).
• Participate in cross-functional agile teams to translate customer requirements into prototype solutions, with opportunities to lead smaller sub-projects.
• Analyze semiconductor testing data (parametric measurements, yield logs) using statistical methods and visualization tools.

• University degree in Data Science, Computer Science, Electrical Engineering, or related field (Master's preferred, Bachelor's with 2+ years' experience accepted).
• 1-3 years of hands-on experience in machine learning, including coursework/practical work with NLP or LLMs.
• Proficiency in Python for data analysis (Pandas, NumPy) and basic ML model development (scikit-learn, PyTorch).
• Familiarity with LLM concepts: transformer architectures, prompt engineering, or text generation techniques.
• Foundational understanding of MLOps practices - version control (Git/DVC), containerization (Docker), and cloud deployment basics.
• Basic Linux/Unix command-line skills and ability to work with Jupyter notebooks or VS Code.
• Strong communication skills in English; ability to document technical work clearly.

This is a plus:

• Exposure to semiconductor testing data or industrial IoT datasets.
• Experience with RAG systems or LLM fine-tuning workflows (LoRA, QLoRA).
• Basic knowledge of electronic measurement principles (oscilloscopes, parametric analyzers).
• Familiarity with Java for integration with existing test platform codebases.
• Elementary German proficiency.