Lead, Real-Time Voice AI Research

Airy; Singapore; On-site; SGD 180,000 - 340,000; Full time

Job summary

Airy is seeking an experienced Head of AI Research (Voice) to steer the development of real-time voice AI technologies. The role leads a small team of researchers, collaborates with engineering, and focuses on pushing the boundaries of speech-to-speech, ASR, and TTS for production-grade systems.

You will drive model training, evaluation, and deployment strategies to ensure low latency, high accuracy, and robust conversation quality across real-time applications.

Qualifications

• Strong track record in AI research and development for speech-to-speech, ASR or TTS.
• Experience at a leading AI research lab or tech company with strong speech AI capabilities.
• Proven experience taking AI research models into real-world production environments.
• Deep knowledge of model training, fine-tuning, optimization, and evaluation.
• Experience building low-latency, real-time voice AI systems is highly preferred.
• Strong technical leadership skills and the ability to lead a small, specialized research team.

Responsibilities

• Lead research and development of advanced voice AI models for natural, real-time conversations.
• Improve speech recognition, speech generation, and AI response quality.
• Develop and optimize AI models to reduce response delays and improve interruption handling.
• Lead model training, fine-tuning, and optimization for real-world production use.
• Build evaluation systems to measure AI accuracy, performance, and conversation quality.
• Lead a small team of AI researchers and collaborate closely with engineering teams.

Skills

AI research

Speech-to-Speech

ASR

TTS

Model deployment

Technical leadership

Education

PhD or equivalent in AI/ML

Tools

Python

PyTorch

TensorFlow

Job description

Airy is seeking an experienced Head of AI Research (Voice) to steer the development of real-time voice AI technologies. The role leads a small team of researchers, collaborates with engineering, and focuses on pushing the boundaries of speech-to-speech, ASR, and TTS for production-grade systems.

You will drive model training, evaluation, and deployment strategies to ensure low latency, high accuracy, and robust conversation quality across real-time applications.