Exciting Opportunity for a Senior ML Engineer!
About the Project:
Join an innovative AI Voice Infrastructure initiative focused on developing state-of-the-art Arabic speech technologies. The project encompasses advanced speech synthesis (TTS), real-time automatic speech recognition (ASR), and end-to-end speech-to-speech conversational systems designed for Arabic dialects, including Gulf, Levantine, Egyptian, and other regional variants.
Key Responsibilities:
- Benchmark TTS and ASR models using Arabic speech evaluation datasets, measuring metrics such as Word Error Rate (WER), naturalness, and dialect coverage.
- Fine-tune and optimize generative architectures for voice cloning and zero-shot speaker adaptation.
- Build, maintain, and improve Arabic-specific data pipelines, including audio sourcing, transcription processing, and diacritization workflows.
- Optimize model inference for production environments through quantization, KV cache tuning, and latency reduction techniques.
- Integrate, test, and evaluate complete speech-to-speech conversational pipelines.
- Conduct research experiments and transform academic advancements into production-ready solutions.
- Collaborate with cross-functional teams to deliver high-performance AI voice products.
Requirements:
- 5+ years of experience in Machine Learning, Deep Learning, or AI Research.
- Strong proficiency in Python, PyTorch, and the Hugging Face ecosystem.
- Hands-on experience training and fine-tuning neural speech models, including TTS, ASR, or Audio Codec architectures.
- Deep understanding of modern speech technologies and architectures such as Whisper, Conformer, HiFi-GAN, Diffusion-based models, or similar frameworks.
- Experience with audio processing techniques, including Voice Activity Detection (VAD), speaker diarization, and neural vocoders.
- Proven ability to implement and validate ideas from research papers through practical experimentation.
- Strong understanding of Arabic NLP challenges, including diacritization (Tashkil) and dialectal variations.
- Experience with inference optimization techniques such as quantization, streaming inference, TensorRT, or related acceleration frameworks.
Nice to Have:
- Experience developing high-performance CUDA kernels.
- Familiarity with speculative decoding techniques for large-scale generative models.
- Prior experience deploying speech AI systems in production environments.
- Contributions to open-source AI, speech, or NLP projects.
Language Requirements: Arabic (Native), English (B1+)
How to Apply:
If you're ready to make an impact as an AI Engineer, apply through LinkedIn by submitting your updated CV.
#Hiring #SeniorMLEngineer #MachineLearning #ArtificialIntelligence #SpeechAI #ASR #TTS #VoiceAI #GenerativeAI #PyTorch #AIJobs #TechJobs
