Knowledge path
Architecture, semantics, tool use, evaluation, security, and production patterns for trustworthy Data Agents.
DataAgentX Tech explains the systems behind reliable agentic data work—from the semantic layer and execution harness to observability, evaluation, and human control.
Starting from the three pillars of DACon 2026, this article explains what unified metadata, open semantic layers, and Ontology each address, why context expires, and which real-world problems should serve as the starting point for construction.
Recalculate payment fee rate changes using two public hard problems from DABstep: how conditions map to MCC codes, how the affected transaction set is determined, why fixed amounts offset in the differential, and how individual differentials are independently verified.
Recalculating hotel brand rankings using two public tasks from InfiAgent-DABench to examine how eligibility thresholds determine the evaluation scope, why denominators cannot be mixed, and how regrouping recalculations can overturn overall conclusions.
Organize 15 Skills with task protocols, work and acceptance chains, six failure tests, and a practical path from one frequent task to a reliable Agent.
Deconstructing Publish Artifact to Sites: prerequisites for publishing, the one-to-one mapping between logical artifacts and sites, the distinction between access policies and editing permissions, and why automated publishing is conditional.
Compare Build Dashboard and Build Report: their audiences, metric models, evidence requirements, acceptance criteria, and delivery conditions.
How Jupyter Notebooks and Visualize Data connect reproducible calculations, chart contracts, and honest presentation in a Data Agent workflow.
Separate checks on source data from independent checks on analytical claims, including grain, missingness, joins, denominators, and causal interpretation.
