FULL-STACK AI
DevMesh AI — Developer Collaboration Platform
Platform for developer team formation and recruitment. Computes a 0–100 Developer Trust Score from five weighted signals and uses OpenAI embeddings to match cofounders and teammates by skill.
Public demo
5 — weighted Trust Score signals
The problem
Developers waste weeks on cold outreach to find compatible cofounders. Traditional job boards show titles, not technical depth — a senior ML engineer and a junior one look identical on paper.
The approach
Built a three-tier platform: React/TypeScript frontend, Node/Express REST API, and a FastAPI AI service. The Trust Score is a server-side weighted composite recomputed on every profile change — GitHub activity 25%, projects 25%, profile completeness 20%, certificates 20%, platform activity 10% — mapped to four tiers. OpenAI embeddings power the matching; Socket.io delivers real-time presence, typing indicators, and read receipts.
The results
Real-time team chat and Kanban boards, an idea board, and a recruiter dashboard ranked by Trust Score. Matching falls back to a keyword-overlap heuristic when the embedding API is unreachable, so the feature returns results rather than an error. No load test was run.
Stack: React 18, TypeScript, Node.js, FastAPI, MongoDB, Socket.io, OpenAI