Author: Jose S
Orchestration in one metaphor
Orchestration is kitchen management for data: schedules, dependencies, and retries. Learn the Airflow and dbt mental models so broken morning numbers point to a clear task, not a…
Prompt patterns for data work
Build data prompts with role, spec, output shape, and refuse rules. Make grain, filters, and calendar explicit so the model cannot hide a wrong query behind fluent SQL.
Quality control and avoiding workslop
Catch workslop before you share: verify numbers, sources, and claims even when the deck looks polished. Use a quality loop by artifact type and a paste-ready review card.
Dashboards vs single slides
Dashboards monitor change; single slides carry one decision. Use the one-idea-per-view rule and a declutter checklist so a busy ops page stops pretending to be a decision document.
Fine-tuning vs RAG
Fine-tuning changes model behavior with examples. RAG retrieves documents into the prompt. Prefer RAG first for company knowledge; fine-tune when style or task format is the real gap.
