Sorin

Language models.Speech technology.

We’re developing Kurdish language and speech technology, and exploring how it can become useful, dependable software.

Language models

In development

Kurdish language models for reading, writing, and working with text.

Text understanding · Language generation

Text to speech

In development

Speech synthesis engines that turn written Kurdish into spoken language.

Voice generation · Language access

Speech to text

In development

Speech recognition engines for transcription and voice-led applications.

Transcription · Voice interfaces

These are development areas, not released products. We’re open to conversations about research, datasets, and practical use cases.

From experiments to useful tools.

Language and context

Build around how people actually write, speak, and use technology.

Evaluation before claims

Test quality, limitations, and failure cases before calling a system ready.

Thoughtful data use

Consider permission, provenance, representation, and the people behind the data.

Explore a research collaboration

Working with language data, speech, or an AI use case? Tell us what you’re exploring.

Start a conversation