.proof-track{height:auto!important}.proof-stage{position:relative!important;height:auto!important;overflow:visible!important}.proof-panel{position:relative!important;min-height:800px;visibility:visible!important;opacity:1!important;pointer-events:auto!important}.proof-panel:not(.personal-panel){padding:100px 6vw!important;display:flex;flex-direction:column;gap:24px}.proof-panel:not(.personal-panel)>*{position:relative!important;inset:auto!important;width:auto!important}.work-media{height:350px!important;width:260px!important}.proof-portrait{visibility:visible!important}.proof-meter{display:none!important}The visual effects are off. Browse the content and project index with normal links.

experiment

Support-ticket tagging

BART and FLAN-T5 comparison on simulated tickets.

Support-ticket tagging: System illustration · not product UI — System illustration · not product UI

01 / The system in contextScope & evidence ↓

Product / evidence

What the evidence supports.

BART and FLAN-T5 comparison on simulated tickets.

An experimental study, not a deployed client system or a validated business outcome.

Open source / public project ↗Have a related problem? Let’s talk ↗

← Return to selected work