How do One Big Table and AI fit together

2024-04-11 · Source: DataExpert.io Newsletter · Field: Technology & Digital — Data Science & Analytics, Artificial Intelligence & Machine Learning · Depth: Intermediate, quick

Summary

Data modeling is evolving significantly with the rise of AI, shifting data engineers towards a new role: the "context engineer." This new role focuses on providing real-time, relevant context to Large Language Models (LLMs) to enable rapid, action-oriented decision-making, moving beyond traditional analytics-focused architectures. The article contrasts three data modeling techniques: One Big Table (OBT), Dimensional Modeling, and AI-native Modeling. While OBT simplifies context retrieval for AI by avoiding complex JOINs, it often provides too much context. Dimensional Modeling offers precise context but relies heavily on text-to-SQL, which LLMs struggle with. AI-native modeling emerges as a question-first approach, aiming to combine OBT's simplicity with Dimensional Modeling's precision by adding an ETL layer to create targeted data marts that map directly to specific questions, thereby optimizing token usage and improving AI accuracy.

Key takeaway

For data engineering leaders building AI-driven systems, your teams should prioritize developing "context engineering" skills, focusing on real-time data provision and AI-native modeling. This shift will enable your LLMs to deliver more accurate, faster, and cost-efficient responses by providing precisely the right context. Evaluate adopting a question-first AI-native modeling approach to optimize data retrieval and minimize token consumption, ensuring your AI applications are both effective and economical.

Key insights

Context engineering is crucial for AI to deliver consistent, low-latency, and cost-effective answers.

Principles

• AI requires the "Goldilocks zone" of context.
• Analysts are a bottleneck in current architectures.
• AI thrives on mapping questions to appropriate datasets.

Method

AI-native modeling introduces an ETL layer to create question-specific data marts on top of OBT models, enabling precise context delivery and reducing token costs for LLMs.

In practice

• Label dimension columns (e.g., dim_name).
• Use column and table comments for text-to-SQL.
• Sharpen system design skills for big data.

Topics

• Context Engineering
• AI-native Modeling
• One Big Table
• Dimensional Modeling
• Large Language Models

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Data Engineer, MLOps Engineer, AI Architect

Related on AIssential

• How Spotify Taught an LLM to Think Like a Senior Data Analyst — LLMs fail from lack of meaningful context, not intelligence; expert-owned context is paramount.
• Presentation: AI Agents to Make Sense of Data at OpenAI — OpenAI's Kepler agent uses advanced AI techniques to automate complex data analysis across vast datasets, improving data accessibility and accuracy.
• I Built an AI Data Engineering Mentor from Scratch: Here’s What Nobody Tells You — AI agent effectiveness is primarily driven by structured context, not just the underlying large language model.
• The Context Problem — AI's "context" as a billing unit conflates capacity with coherence, leading to inflated costs and degraded model performance.
• Luis Serrano + Jay Alammar + Josh Starmer Q&A Livestream!!! — Effective AI education combines clear explanations, relatable analogies, and strong visual communication.
• We Don’t Have an AI Problem. We Have a Context Problem. (And MCP Exposes It) — Unstructured context, not model limitations, is the primary bottleneck in scaling and maintaining LLM-based systems.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by DataExpert.io Newsletter.