Tech Desk 
Apr 09, 2025
Amid fierce global competition in artificial intelligence, Pakistan appears to be carving out its fair share as a 23-year-old developer from Hyderabad, Sindh, has successfully created the first-ever AI tools to support the Sindhi language.
Fahad Maqsood Qazi has designed the AI model for both text-to-speech (TTS) and speech-to-text (STT) systems for Sindhi, which is spoken by nearly 40 million people worldwide.
As reported by ProPakistani, he started working on the project in 2023 while developing an AI-driven dubbing system. What prompted him to indulge in the project was the absence of even basic AI tools for Sindhi.
Qazi is said to have gathered hours of transcribed audio content from YouTube, audiobooks, and news reports in order to supplement it with data from Mozilla’s Common Voice project, which recently had started supporting Sindhi.
Qazi had completely developed initial versions of TTS and STT tools by January 2024 alongside a language tokenizer —an essential part for machine learning models.
His AI tools have emerged into the limelight for catering to Sindhi-speaking communities and closing the digital accessibility gap.
In March this year, the Sindhi language AI tools were made publicly available on HuggingFace, giving access to global developers.
