​Right-Sizing Your Analytics Architecture: From Laptop to Lakehouse

Build too small and you hit scaling walls. Build too big and you waste time and money on complexity you don't need yet. This talk offers a practical middle path: analytics systems that start small and grow to enterprise scale without a rewrite.

​

Aaron Spiegel, Derek Moore, & Elizabeth Christensen will walk through a real-world pipeline that begins on a single machine using open-source tools like Polars, Postgres, and Duck

DB, then moves to distributed platforms like Databricks or Snowflake as data volumes grow. The key is making smart choices about data formats and storage from the start, using open table formats like Apache Iceberg, so you can postpone big scaling decisions without creating technical debt.

​

You'll learn:

* ​When a single machine is enough and when you need distributed processing

* ​How open table formats let you switch processing engines using the same data

* ​How GPU acceleration can stretch single-node performance further

* ​A practical framework for architecture