David Vinkenroye
Menu

Emineon OS

An AI-native ATS and CRM for recruiters, with CV parsing, matching and a copilot.

Built by in Geneva.

The story

I built Emineon OS because I saw recruiters, including me, copy the same candidate into five different tools before lunch. I wanted to keep candidates and client work in one place. It is the software Emineon works with.

It parses CVs and LinkedIn profiles into structured profiles and matches candidates to jobs with hybrid and semantic search on pgvector. It creates branded competence files and presentations as PDF, DOCX and PPTX, and an agentic copilot with tools drafts emails and job descriptions. Voice endpoints handle transcription.

Other agents can query it through a read-only MCP server with OAuth resource discovery. There is a Chrome extension for LinkedIn and an add-in for Outlook. It runs in the cloud or self-hosted with Docker Compose, Postgres, Redis, MinIO and optional local models through Ollama. It is still in beta.

What was built

  • CV and LinkedIn parsing into structured candidate profiles
  • Candidate-job matching with hybrid and semantic search on pgvector
  • Branded competence files and presentations in PDF, DOCX and PPTX
  • An agentic copilot with tools, plus speech-to-text and text-to-speech
  • A read-only MCP server with OAuth resource discovery
  • A Chrome companion for LinkedIn and an Outlook add-in
  • Self-hostable, with optional local models

Stack

  • Next.js
  • TypeScript
  • Prisma
  • PostgreSQL + pgvector
  • Redis / BullMQ
  • OpenAI Agents SDK
  • Ollama
  • MCP
  • Clerk
  • MinIO / S3
  • Docker Compose
  • Vitest

Themes

  • AI agents

    I wanted a personal AI assistant that can do everything online for me, and it can even make phone calls. I build them with whichever models and tools fit best: Claude, OpenAI, Gemini, open models and plenty of plain code.

  • MCP and tool use

    I use the Model Context Protocol to connect AI assistants to real things: a camera, a photo library, my own archive and Emineon OS, my recruiting tool. I keep each server small with only a few tools, and read-only where that matters.

  • Retrieval and grounding

    I want models to answer from the original records, so they read them when the question is asked. I measure recall against real questions. In Lodestar, a camera setting only counts once the camera reads it back.

  • Self-hosting and data sovereignty

    Personal data stays on hardware its owner controls. For me that's a NAS behind a tunnel. Each Neonduck customer gets their own isolated server, and if they delete the account, the server is deleted too.

  • Tech people

    A lot of my work is about people in tech. Emineon and Haystack help companies hire tech talent. I built Emineon OS, the recruiting tool I wanted while doing that work, and Romandy CTO brings tech leaders together.

  • Emineon OS is the software product of Emineon.
  • DbrainSelf-hosted memory for my AI assistants: verbatim records on a home NAS, served over MCP.
  • LodestarAn iPhone and iPad app that sets up a Sony camera from one sentence and takes the shot.
  • NeonduckA personal AI assistant that does everything online for you and can even make calls.