PhD Position Decentralized and Trustworthy AI Pipelines (AI)

Location: Based at hirify.global in Delft, the Netherlands; relocation to the Netherlands may be required

Salary: €3,204–€4,051 gross per month, increasing from the first to the fourth year

Company

hirify.global is an international technical university conducting research and education across engineering, science, design, and digital society.

What you will do

• Design decentralized and federated AI training and inference pipelines that combine real and synthetic data across organisational boundaries without a central aggregator.
• Develop resilience against Byzantine participants and poisoning of synthetic data.
• Measure membership inference, reconstruction risks, privacy leakage, and the utility costs of differential privacy.
• Study distributed LLM-based systems, including model-failure correlation, quorum mechanisms, and semantic agreement.
• Build benchmarks, cross-validation schemes, and reporting protocols covering accuracy, robustness, bias, generalisation, and privacy.
• Publish research, release reproducible open-source software, collaborate with European research groups and public authorities, and contribute up to 15% to teaching and supervision.

Requirements

• MSc completed or near completion in Computer Science, Data Science, Artificial Intelligence, Electrical Engineering, or a closely related field.
• Strong foundation in machine learning and/or distributed systems, with strong programming skills in Python.
• Hands-on experience with a modern deep learning framework such as PyTorch and Linux-based GPU/HPC clusters.
• Interest in trustworthy AI, including privacy, robustness, adversarial behaviour, evaluation methodology, and reproducibility.
• Analytical skills, research independence, and the ability to maintain a four-year research agenda while meeting project deadlines.
• Excellent written and spoken English is required; non-native speakers without an English-taught degree must meet hirify.global requirements such as TOEFL iBT 90 or IELTS 6.5 overall.

Nice to have

• Familiarity with federated or decentralized learning, fault tolerance, generative models, privacy-preserving machine learning, or LLM-based systems.
• Experience explaining technical work to public administrations and other non-academic stakeholders.
• Links to open-source code repositories.

Culture & Benefits

• Four-year doctoral employment in principle, structured as an initial 1.5-year contract followed by a 2.5-year extension subject to a go/no-go assessment and performance requirements.
• Enrollment in the hirify.global Graduate School and participation in its doctoral education programme.
• 8% holiday allowance and an 8.3% end-of-year bonus.
• Customisable compensation package, health-insurance discounts, monthly work-cost contribution, and flexible work schedules.
• Access to DelftBlue and DAIC GPU clusters and collaboration across a consortium of more than thirty partners.
• Relocation information, settlement events, and a dual-career programme for accompanying partners.

Hiring process

• Apply online by 24 September 2026 with a cover letter, CV, academic transcripts, MSc thesis if applicable, and proof of English proficiency if applicable.
• Recruitment includes a knowledge-security risk assessment during the final stages.