PyIceberg is Python's native implementation for Apache Iceberg, providing lightweight table operations without JVM dependencies. This hands-on workshop provides a technical dive into PyIceberg's architecture, focusing on read and write paths and performance optimization. Participants work through practical exercises covering read path architecture (metadata processing, scan planning, predicate pushdown, partition pruning, time travel), write path architecture (table creation, append/overwrite operations, partitioning strategies, ACID guarantees, conflict resolution), and advanced topics (schema evolution, partition evolution, snapshot management, catalog integration). Prerequisites: Basic Python knowledge and data lake familiarity. Iceberg experience helpful but not required. Attendees bring laptops; setup instructions provided in advance.