about

7,000+ languages exist. AI works for maybe 100 of them. I’m working to fix that.

I build speech and language tech for languages the industry ignores. Not because it’s easy (it’s not), but because every language deserves a voice in the AI age.

PhD in computational linguistics, 10 years getting my hands dirty with messy data, reluctant communities, and zero-resource scenarios. I’ve built the first neural voices for Catalan, Kanuri, Marma and Judeo-Spanish, led community-powered data collection campaigns for Aranese and Tamazight, and worked on innovative projects across 15+ languages from Africa, Asia, including those in endangered category.

What I do: turn “impossible” into working prototypes. ASR for impaired speech. MT for endangered languages. Chatbots that engage, not just chatter. I work with NGOs, researchers, and language communities who need someone technical but also gets why your grandmother’s language matters.

Father. Based in Barcelona. Born in Turkey, lived in Germany. Speak Turkish, English, Spanish, some Catalan. Enjoy camping, biking, drumming, and conversations about how to build tech that actually serves people—not the other way around.