- Zürich, Zurich, Switzerland
- On-site
- Full-time · Senior
- 7+ years
Design and build systems to capture tacit knowledge from enterprise documents, tickets, and conversations, modeling it as a governed knowledge graph for AI agents.
Responsibilities
- Research and build ways to capture decisions, expertise and context from enterprise documents, tickets and conversations, and model them as a knowledge graph with evidence, provenance, versioned assertions, access controls, and human review
- Follow research on knowledge graphs, LLM-based extraction and agent memory; build prototypes, design benchmarks and evaluations, and turn successful experiments into production features
- Make organisational memory available to LLM agents through typed APIs and MCP, and measure the quality of the answers they give
- Improve query parsing, structured search, typeahead, faceted navigation and search relevance measurement in enterprise search
- Develop Python services and typed APIs, maintain production frontend code, automated testing and CI, and support security reviews
Requirements
- University degree in a relevant field
- Nice to have: PROV-O
- Nice to have: SKOS
- Nice to have: Terraform
- Nice to have: Ansible/AWX