DATAmundi

DATAmundi is a Belgian AI training-data company founded in 2007 by Gert Van Assche and his wife in Linter, Belgium. Originally a language-services boutique, the company moved into AI training data well before the LLM era and built a reputation for high-quality multilingual datasets across speech, text, image, and video. In December 2021 DATAmundi was acquired by Polish language-services group Summa Linguae Technologies, which subsequently consolidated its AI-data offering under the DATAmundi brand and relaunched as DATAmundi.ai, "data of the world". The company supplies ethically-sourced, bias-aware training data to AI labs and enterprises through a hybrid model: a ~35-person core team in Belgium directs project delivery against a broader contributor / language-expert network mobilised per engagement. Services include speech data collection and transcription, multilingual text annotation, image and video labelling, MT post-editing, search relevance evaluation, and human-in-the-loop QA. Specialism is European and rarer-language coverage where larger competitors struggle to staff quality contributors.

DATAmundi hires across two tracks. Core team (Belgium / Summa Linguae offices in Poland and elsewhere) covers project management, linguistics, quality, data engineering, sales, and account management; these are local-employment roles typically based in Belgium or other EU Summa Linguae hubs. Contributor track covers freelance / per-project annotators, transcribers, voice talent, evaluators, and language experts, recruited per engagement against specific language pairs and domains (especially European and rare languages). Contributor work is fully remote with a country / native-language match per project. Applications go through the careers section of datamundi.ai and recruitment-page postings on the company's LinkedIn.