Performance, score integrity, and usability work inside a live NCAA gymnastics product where match-window traffic and small scoring errors directly affect trust.
The live scoring product served a seasonal audience of more than 150,000, with usage concentrated around live meet windows. People relied on current scores, rankings, athlete inclusion, and clear presentation while competitions were underway.
This was established production software with live data and an existing audience—not a student recreation or isolated prototype. The audience figure describes the scale of the product, not traffic caused by my individual changes.
Contribution boundary
I worked on the meet-viewer branch and its user-facing behavior. I did not build the entire Virtius platform, scoring operation, or underlying competition infrastructure.
02 / Contribution
Three product priorities
PerformanceLoad the live view better
Improved loading behavior, API data handling, caching decisions, and client/server boundaries for match-driven use.
Score integrityProtect the displayed result
Handled precision, three-versus-four-decimal display, exhibition athletes, rankings, and related scoring edge cases.
Usability and designMake results easier to follow
Made presentation and interaction changes that improved how users read and navigate the live viewer.
Production interfacePublic live-system screenshot
03 / System
Request to live result
01Meet and scoring data
Competition, team, athlete, and score data enter the established product flow.
02GraphQL and application layer
Requests, caching, data boundaries, and rendering decisions shape viewer state.
03Live meet interface
Scores, precision, rankings, filters, and current results are presented to users.
04 / Verification
Changes had to survive live data
I studied the request flow used by the earlier Virtius experience rather than guessing at its contract. Playwright checks, visual comparisons, and expected scoring cases helped verify performance and interface changes without weakening score behavior.
Playwright checks, scoring display cases, and visual comparisons.
05 / Result
A faster, clearer, more trustworthy viewer
The work improved how the live product loaded, protected score presentation across edge cases, and made results easier to use. The strongest public proof is the product itself; exact causal traffic, conversion, and private engineering metrics are intentionally not claimed.
