Track

AI & AI Systems Engineering

A model on its own is not a product. An AI system also needs the right data, retrieval, tools, evaluation, security and day-to-day operations. This track treats AI as an engineering discipline, built on the same data platform as the other tracks: the warehouse, the semantic model and Airflow.

New track · in progress LLMs · RAG · Vector search · Agents · Evaluation · LLMOps Level: intermediate → advanced

Tutorials

Tutorial · AI systems

AI Systems Engineering: From Data Platform to Production AI

From how LLMs behave to embeddings, retrieval over warehouse data, text-to-SQL on a semantic layer, tools and agents, then evaluation, guardrails, LLMOps and orchestrating AI pipelines with Airflow.

How it connects to the other tracks

Data Warehousing provides the governed, tested data that AI answers are grounded in.
Business Intelligence provides the trusted measures that text-to-SQL must use, not bypass.
dbt and Airflow keep embeddings, batch inference and evaluation runs fresh and repeatable.
Statistics is the basis for measuring whether an AI system is actually getting better.