Designs by Duhart · work inventory
Scored on how much real distributed-systems engineering each one carries, not on how impressive it looks.
Each project is rated 0–5 per signal. Ratings are a hiring judgment, not a measurement, treat differences under about five points as ties.
The core distributed-systems work.
Within 5 points of 120, and that is noise. GlobalChat edges the rubric on production evidence and breadth. The Activity Graph leads anyway because it is the canonical systems-design question with every trade-off already written down. Open with it; bring GlobalChat as proof you have carried something this size in production.
Strong support, infrastructure, ML platform, media.
Real systems work, narrower in scope.
Breadth and depth-of-craft. Mention, don’t lead.