You join a motivated and interdisciplinary team working on the Zurich Legal LLM project at the Faculty of Law of the University of Zurich. You help develop and test AI systems for legal applications, gaining hands-on experience in applied machine learning, NLP, and AI governance. This role offers flexible hours that fit with your Master's studies and the chance to work partly from home.
Your tasks
- Support the design, implementation, and documentation of a controlled RAG pipeline for Swiss legal materials
- Prepare, clean, and structure legal teaching materials for retrieval-based use
- Implement document ingestion, chunking, embedding, retrieval, and citation-tracking workflows
- Test and compare open-weights LLMs and commercial systems under controlled conditions
- Implement guardrails against content leakage, excessive quotation, and extraction-style prompting
- Help develop a benchmark suite for doctrinal Q&A, exam-style legal reasoning, and structured objection-response drafts
- Contribute to automatic evaluation workflows, such as faithfulness checks, citation precision, retrieval quality, and structure compliance
- Maintain technical documentation, experiment logs, and reproducible code
- Assist with developing an internal teaching-oriented prototype, including exam-question generation and response-drafting workflows
- Collaborate with legal researchers, student assistants, and faculty members in an interdisciplinary project
What you bring
- Ongoing or recently completed Master's studies in computer science, data science, computational linguistics, artificial intelligence, software engineering, or a related quantitative field
- Good programming skills in Python
- Familiarity with machine learning, NLP, or information retrieval
- Interest in Large Language Models, Retrieval-Augmented Generation, and applied AI systems
- Ability to work carefully with structured and unstructured text data
- Good written and spoken English
- Motivation to work in an interdisciplinary environment at the intersection of law, data science, and AI governance
- The following are assets, but not required: experience with RAG frameworks, vector databases, graph RAGs, embedding models, LLM APIs, open-weights LLMs, local or controlled model deployment, tools like LangChain, Hugging Face, Ollama, LLM output evaluation, benchmarking, prompt engineering, hallucination/faithfulness testing, document processing workflows (PDF, DOCX, Markdown, OCR), basic knowledge of German, interest in legal data science, empirical legal research, or AI governance
- Strong motivation, reliability, attention to detail, clear documentation habits, and ability to work independently on technical tasks are valued
What we offer
- Challenging applied research position in a motivated and interdisciplinary team
- Opportunity to contribute to a strategic AI project at the Faculty of Law
- Hands-on experience with RAG systems, LLM evaluation, legal NLP, and AI governance
- Close collaboration with legal scholars and data scientists
- Chance to contribute to academic publications and internal prototypes
- Flexible working hours compatible with Master's studies
- Possibility of working partly from home
- Experience building AI infrastructure for research and teaching use cases
- Diversity and inclusion are important values
Bist du Teil dieses Unternehmens?
Dieses Unternehmensprofil wurde automatisch erstellt. Wenn du für Universität Zürich arbeitest, kannst du das Profil jetzt übernehmen und verifizieren – kostenlos und in wenigen Minuten.
Über uns
The University of Zurich is Switzerland's largest university, offering an inspiring environment for cutting-edge research and top-class education. With about 10,000 employees, it provides diverse opportunities in a wide range of fields, supporting talent, innovation, and inclusion.