← Back to AI PillarsPillar 6
A skill vocabulary employers can actually search.
Cost per call:₹0 · In-house inference
Raw skill labels on profiles are noise — "React", "React.js", "ReactJS", "Reactjs" are the same to humans and four different strings to a database. The Skill Taxonomy Engine maps every submission to a standardised tag tree so employer search, leaderboards, and portfolio aggregation all speak the same language.
Public surface
GET /skills/{skill}Model routing
In-house specialized inference
Cost envelope
~₹0 per submission mapped
What gets mapped
- • Domain tags — SWE, Data, Design, PM, Marketing, Finance, Content, HR, Ops, BD
- • Framework tags — React, Next.js, Django, Figma, Tableau, Airtable…
- • Competency tags — research rigor, visual craft, systems thinking, documentation, stakeholder comms…
- • Level bands — entry-level, intermediate, advanced — derived from score × difficulty × consistency
Every public /p/<username> profile and /skills/<skill> landing page is rendered from this taxonomy. Synonyms collapse automatically. The vocabulary evolves quarterly, versioned.