IS 455: Database Design & Prototyping

Spring 2026, University of Illinois Urbana-Champaign

IS 455 teaches students to build reliable data systems from messy real-world inputs. The course starts with relational databases and SQL, then moves through data auditing, schema design, reproducible pipelines, semi-structured data, vector search, retrieval-augmented generation, graph-based retrieval, and monitoring.

The course is designed around a practical principle: a data pipeline should be inspectable, reproducible, and robust to edge cases. Students work with Python, Jupyter, DuckDB, SQL, and modern retrieval workflows while learning how to document assumptions and test whether a system still works outside the happy path.

Core themes:

• database design for real data rather than toy tables;
• reproducible analysis and explicit data lineage;
• joins, keys, granularity, and aggregation risk;
• modern retrieval systems, including vector search and RAG;
• monitoring, reconciliation, and failure analysis.