BuildDiff
A deterministic comparison platform for regulated web content
- Role
- AI Solutions Engineer — designed and built solo
- Timeline
- 2025 — present, in production
- Users
- Copyeditors, QA validators, front-end developers
- Outcome
- 90% fewer missed editorial edits · $300K annual savings
Before
Copyeditors, QA validators, and front-end developers at Medscape were reviewing web builds by hand, checking text, styling, markup, images, layout, and tracking links across versions. A missed editorial edit in a medically regulated environment is a compliance problem, not a typo.
Architecture
I designed and built BuildDiff to replace the legacy tooling and the manual passes around it. It compares live URLs and HTML files across six layers and presents results through Visual, Source Code, and Overlay views, with export to marked-up PDF and post-merge HTML. It runs in production today.
The key design decision
The comparison engine is fully deterministic. In a medical-legal workflow, results have to be reproducible and explainable before anyone will act on them. The AI layer, LangGraph triage and plain-language summarization, sits on top of that engine, read-only, and never as the source of truth.
What went wrong
Getting BuildDiff live inside WebMD's older DevOps environment was the constraint I underestimated. The deterministic engine was finished well before it had somewhere to run.
Results
What I'd do differently
I would sequence the hosting work first next time, and validate the deployment path before building the second half of the diff layers.