Mandatory Internship Automating High-Speed PCB Material Databases Using AI and Data Engineering

Modern high-speed electronic systems rely on accurate material parameters such as dielectric constant (Dk) and loss tangent (Df). These parameters are typically published by suppliers in large technical datasheets and lineup documents. Extracting and maintaining this information in a structured form for engineering use is currently a manual and time-consuming process.

In this internship you will develop an AI-assisted pipeline that converts technical documents into structured engineering knowledge. The project focuses on applying machine learning and document understanding techniques to automatically identify, extract, validate, and structure material parameters from supplier documentation. The goal is to build a reproducible data pipeline that transforms unstructured documents into a structured knowledge base used in engineering simulations.

• During your assignment you will develop a document understanding pipeline to identify relevant parameter tables in technical documents.
• You will implement machine learning / LLM-based extraction of material parameters and metadata and design a structured representation of material knowledge including provenance and validation rules.
• Furthermore, you will build automated workflows that convert extracted data into simulation-ready datasets.
• For extracted parameters you will implement quality checks and anomaly detection. You will also explore methods to improve extraction accuracy using ML techniques (prompting, classification, model evaluation).
• Additionally, you will create tools that allow engineers to review and approve extracted information efficiently.
• Finally, you will document the workflow and evaluate the performance of the developed system.