FROM RESEARCH TO AI ACCELERATION: UNLOCKING HIGH-PERFORMANCE COMPUTING FOR ACADEMIC INNOVATION »

Dr. Abdelwahab Heba dirige également Data Center IA Algerie – SARL AI GRID SOL, il propose des solutions efficaces avec des cluster de calcul IA.

Biographie

Abdelwahab Heba is a Senior Applied/Research Scientist specializing in speech processing, machine learning, and AI-driven language modeling. His research has significantly contributed to end-to-end speech recognition, speech translation, and multimodal AI.

He completed his PhD in France, focusing on speech sequence modeling, and later conducted research at MILA under the supervision of Mirco Ravanelli and Yoshua Bengio. Abdelwahab played a key role in developing SpeechBrain (https://speechbrain.github.io), an open-source toolkit for end-to-end speech modeling, which has become widely adopted in academia and industry.

At Microsoft, Abdelwahab has driven innovation in streamable RNN-T models, multilingual acoustic modeling, and efficient AI training techniques through custom kernel optimizations. His work directly powers Microsoft’s AI speech solutions and Azure AI services.

As Former Chief Scientist/Investor at Gladia, Lipitt, FluentX, Zaion, Viadialog startups, he leads advancements in Live Speech Translation, Voice Cloning, and Video Dubbing, shaping real-time multilingual communication. He has also played a pivotal role in speech-to-speech translation for call centers and continues to push the boundaries of AI in speech applications.

Abdelwahab’s recent interests include multimodal large language models (LLMs) and AI Training/Inference optimization, as part of this vision, he is spearheading a new initiative to establish a cutting-edge AI Data Center in Algeria, aiming to drive innovation in computational AI, democratize access to high-performance computing, and foster a local ecosystem for groundbreaking research.