EXPI — Expert network platform · AI search

International platform where companies and organisations find, contract and work with experts, with natural-language AI search over profiles.

Role: Team lead and software architect @ WeDo; AI search and backend

Client: International expert network

Team: 3–4 developers

Period: 2023–2026

Companies looking for niche expertise rarely know the right keywords, and once they find someone, the collaboration scatters across email, calls and spreadsheets. EXPI set out to cover the whole cycle in one place: find an expert or partner by describing the need in plain language, agree terms, work together and get invoiced. The same platform also had to run as white-label deployments for partner initiatives.

We built a Laravel API with a Nuxt 3 and Pinia frontend. Profiles are turned into vector embeddings stored in MySQL, so a query such as “machinery engineering company in the EU” returns semantically close experts, partners, projects or funding. Each collaboration gets a workspace with chat, offers, time reports, invoices and activity logs, plus built-in audio and video calls. Referral rewards, an admin dashboard and realtime WebSockets complete the product. I led the team and implemented the AI search layer.

AI search: natural-language queries over vector embeddings of profiles

Search for partners, services, projects and funding with ratings and rates

Project workspace: chat, offers, time reports, invoices and logs

Built-in audio and video meetings for up to six participants

Cause pages with donations; referral rewards and admin panel

Built 2023–2026; also runs as white-label hubs, including one for Ukraine’s reconstruction

Laravel 9, Nuxt 3, Vue 3 + Pinia, MySQL + vector search, OpenAI embeddings, Laravel WebSockets, AWS S3, Docker, SonarQube