Chris Davis · data & analytics engineer
I help make your dbt project AI-ready.
You already built the warehouse and dbt project for dashboards, reporting, and internal analysis. That part works.
When teams start layering on copilots, analyst agents, or AI-facing analytics features, they usually discover the same thing: the project is useful for humans, but not explicit enough for machines.
I help data and engineering teams tighten definitions, expose business context, and produce artifacts that make AI over warehouse data more trustworthy.
What I help with
Focused help for teams trying to make warehouse-backed AI actually usable
I review the models, joins, documentation, tests, and business definitions behind your target use case and show where an AI feature is likely to misinterpret the data.
I help make implicit business meaning explicit so your team can support copilots, analyst agents, and other AI consumption patterns without relying on guesswork.
The output is not a black box. I deliver structured artifacts and project updates your team can maintain, extend, and plug into the AI stack you choose.
The offer
The AI Context Sprint
A focused three-week engagement for teams with a working dbt project that is not yet ready to support AI-facing analytics use cases.
Week 1 — Audit
I ingest your manifest.json and catalog.json, score the project for AI-readiness, and identify where missing context, brittle joins, or unclear business meaning are likely to cause bad answers.
Week 2 — Remediation
I work inside your existing dbt project to tighten definitions, annotate relationships, map business terms, and reduce the highest-risk gaps for your target AI workflow.
Week 3 — Handoff
You get an AI-readiness report, updated project context, and a structured semantic package your team can use in a copilot, RAG pipeline, or agent flow, plus a walkthrough session so nothing gets lost.
A good fit
- You already have a functioning warehouse and dbt project.
- You have a real AI initiative in flight or on the roadmap.
- You want a focused project that improves readiness without replacing your stack.
Not a good fit
- You are still standing up your first warehouse or dbt project.
- You want general-purpose data consulting or staff augmentation.
- You need someone to permanently own the outcome for your team.
Background
A data engineer’s view of why AI projects break
I am a staff-level data and analytics leader with 10+ years of experience building data systems from zero and improving them through growth, transition, and operational pressure.
My work has spanned analytics engineering, semantic modeling, reporting infrastructure, forecasting support, and product-minded data systems. I have built stacks from scratch, stabilized messy environments, and supported executive reporting in companies ranging from startups to larger organizations.
I tend to think about data in three layers: getting the data into shape, defining the metrics and meaning clearly, and delivering it in a form people can actually use. AI pushes hardest on that middle layer. That is where most teams have more implicit context than they realize.
Built serverless ELT pipelines and analytics products using Python, Modal, and DuckDB.
Advanced SQL, Python, dbt, warehouse design, data modeling, and product-facing analytics workflows.
From first data hire work to staff-level architecture and cross-functional leadership.
Helping teams prepare real data systems for reliable AI consumption instead of improvising on top of ambiguity.
Writing
Posts
Notes, essays, and practical thinking. Each post can link directly to a markdown file in GitHub.
Markdown post
How to do Data
A framework for thinking about DataOps, MetricsOps, ProductOps, and how data teams become more than dashboard factories.
Read on GitHub →
Markdown post
Another note goes here
Replace this with a direct link to a markdown file in your GitHub repo when you are ready.
Read on GitHub →
Contact
Need help making your dbt project AI-ready?
Send a note with a little context about your dbt setup, your AI initiative, and where the uncertainty is. If it looks like a fit, I’ll suggest a focused next step.
Email: christophergdavis@gmail.com
LinkedIn: linkedin.com/in/thedatadavis
GitHub: github.com/thedatadavis