Migration IQ 2.0

The problem

A database migration can look perfectly fine in Dev, QA, and Perf — and behave very differently in Production, where the same operation may encounter tables with billions of rows. By then, finding out that a migration is risky is a little late.

So we asked a simple question: can we use AI and real performance evidence to identify that risk before release night?

The approach

Migration IQ 2.0 connects to an Oracle performance environment, examines actual Flyway execution history together with database metadata, identifies migrations that deserve attention, and uses AI to turn the technical evidence into something developers, DBAs, and release teams can act on.

The decision that mattered

The longer-term idea is deliberately simple: PASS → REVIEW → STOP.

Not AI making the release decision. AI helping humans make a better-informed one. That’s probably the part of the project I value most.

Where it stands

A team hackathon project, built over the Sage Fall 2026 Hackathon, where it received the “Do The Right Thing” Award. The award is appreciated. The more interesting part is seeing a recurring operational problem become a working idea that could help prevent the problem in the first place.

What I took from it

There is a lot of excitement around increasingly sophisticated AI systems and agents. But some of the most useful applications I’ve found start much closer to everyday engineering: where are we repeatedly making an important decision with incomplete information — and can AI help bring the right evidence together earlier?

This was one of those problems.