PhD: Agentic AI for Autonomous Materials Discovery

DIFFER; Netherlands; On-site; EUR 35,000 - 45,000; Full time

Job summary

DIFFER Research in Eindhoven seeks a PhD candidate to join the AMD group, building an agentic AI workflow that converts AI-suggested alloy compositions into executable synthesis and analysis plans. You will work with researchers and industry partners, both in lab automation and data pipelines, to create structured evidence for model development.

The role blends digital workflow development with substantial bench work—about one third digital, two thirds experimental—while training and support are

Qualifications

• Master's degree in CS/AI/related field.
• Strong Python skills and software-engineering practice (Git, testing, reproducible code).
• Experience or interest in agentic AI systems, data pipelines or tool use.
• Willingness to work in a lab and collaborate across disciplines.
• Nice to have lab automation or image/spectral data analysis.

Responsibilities

• Design and run an agentic AI-driven experimental workflow translating AI-proposed compositions into synthesis/characterization plans.
• Carry out high-throughput synthesis of material libraries.
• Characterize samples with micro-XRF and SEM; develop Python pipelines to extract metrics.
• Develop and maintain a FAIR-aligned dataset including realizability labels.
• Connect AI agents to data, planning and reporting tools for traceability.
• Write well-tested software, contribute to publications and complete a PhD within four years.

Skills

Python

Agentic AI

Git

Testing

Reproducible code

Lab automation

English

Multidisciplinary teamwork

Education

Master's degree in computer science

Master's degree in AI/data science

Tools

Python pipelines

Job description

DIFFER Research in Eindhoven seeks a PhD candidate to join the AMD group, building an agentic AI workflow that converts AI-suggested alloy compositions into executable synthesis and analysis plans. You will work with researchers and industry partners, both in lab automation and data pipelines, to create structured evidence for model development.

The role blends digital workflow development with substantial bench work—about one third digital, two thirds experimental—while training and support are