Co-Founder & CEO at Addis AI.
Building voice-first AI infrastructure for African languages.
Mostly working on:
- Speech recognition
- Text-to-speech
- Large language models
- Low-resource language modeling
- African-language datasets
- Model training and evaluation
- AI infrastructure
- Developer tools
I publish most of my open-source models and datasets on Hugging Face.
Some of the work there includes:
- Amharic ASR models
- Amharic language models
- Amharic text and instruction datasets
- Afaan Oromo speech and TTS datasets
Voice, speech, and language models available through APIs and official SDKs.
Python PyTorch Transformers Hugging Face NLP ASR TTS LLMs
Website · Hugging Face · LinkedIn · Addis AI