Building AI-agentic code review into a Fortune 500 data platform
How a multi-agent MCP framework and a hierarchical skill library let AI agents review production code and pick up backlog work directly, alongside a senior architect.
The situation
At the scale this Fortune 500 data platform reached — well over 4,000 dbt models across Snowflake, Databricks, and Azure — manual code review had become the bottleneck. Reviewer bandwidth couldn’t keep up with the rate of change, and quality gates were the first casualty whenever timelines got tight.
What changed
- Built an AI-agentic PR review framework using the Model Context Protocol (MCP), running as one of two mandatory reviewers on every change alongside a human.
- Architected a multi-agent MCP framework backed by a hierarchical knowledge base — a ‘skill’ library — that lets agents pick up, scope, and deliver backlog items autonomously.
- Led ongoing ontology design work so the platform’s data is natively consumable by AI agents, not just BI tools and dashboards.
Why it matters
Most ‘AI for data platforms’ conversations stop at a chatbot bolted onto a warehouse. The more durable pattern is agents that participate in the actual engineering workflow — reviewing code, picking up backlog, reasoning over a shared ontology — without lowering the bar a human reviewer would hold.
Client details anonymized at the customer’s request. Figures and scope are accurate to the engagement.