You'll design and build the multi-agent system behind automated claim generation and processing from document ingestion and structured extraction through reasoning, validation, and steps where experts stay in control.
That includes the document pipelines underneath: structure-preserving parsing, clause-level chunking, and hybrid RAG. And the extraction schemas (Pydantic, JSON Schema) that keep results deterministic, and refined continuously with legal and architectural experts, based on argument mining principles.
You'll work in an agile and results-driven environment, taking the product from a working MVP to a scalable enterprise-grade solution, in close collaboration with Legal, Contract Management, Compliance and IT/Security, and coordinating external technology partners.
Requirements
- A completed or advanced degree in Computer Science, Computational Linguistics, AI or a related field ideally with an NLP and/or argument mining focus
- Hands-on experience building LLM-based systems and taking them into production, not just prototypes
- Solid track record in structured information extraction and working with frontier models
- Good command of modern document processing technologies and methods
- Experience with orchestration frameworks (e.g. LangGraph) and designing multi-agent workflows
- A firm grasp of compliance, data protection and governance in regulated environments (GDPR, revised Swiss FADP, data residency)
- Ability to communicate complex technical concepts clearly and appropriately to different target audiences
- High level of personal accountability, pragmatic execution skills, and readiness to deliver measurable results through short iteration cycles
- Language: English, French and/or German is a plus
Benefits
- Start: ASAP
- Duration: 2-3 months
- Workload: 70-100%
- Salary: 45+ CHF/h
Skills Required
- Completed or advanced degree in Computer Science, Computational Linguistics, AI, or a related field
- NLP and/or argument mining focus
- Hands-on experience building and deploying LLM-based systems in production
- Experience with structured information extraction and frontier models
- Strong knowledge of modern document processing technologies and methods
- Experience with orchestration frameworks such as LangGraph and multi-agent workflow design
- Understanding of compliance, data protection, and governance in regulated environments, including GDPR, revised Swiss FADP, and data residency
- Ability to communicate complex technical concepts clearly to different audiences
- Strong personal accountability and pragmatic execution skills
- Ability to deliver measurable results through short iteration cycles
- English proficiency
- French and/or German language skills
What We Do
ETH juniors connects academic talent with industry by organizing and delivering projects for companies ranging from nationally operating small and medium-sized enterprises to international corporations. It provides project-based consulting and recruits students and young professionals for software, IT, energy trading, AI, data engineering, and related assignments, helping organizations access specialized expertise while giving participants practical experience across diverse business and technology needs.








