Company:
IF Recruitment Ltd
Location: Cambridge
Closing Date: 27/11/2024
Hours: Full Time
Type: Permanent
Job Requirements / Description
We're seeking an exceptional AI Voice Synthesis Developer to join an innovative start-up. The ideal candidate will combine deep technical expertise in text-to-speech (TTS) systems with a passion for creating efficient, production-ready solutions that push the boundaries of what's possible in voice synthesis.
Key Responsibilities
- Design and implement low-latency TTS systems optimised for minimal computing resources
- Develop and optimise AI models for Real Time voice synthesis
- Create efficient architectures that balance quality, speed, and resource utilisation
- Collaborate with team members to integrate voice synthesis capabilities into our products
- Research and implement state-of-the-art techniques in speech synthesis
- Contribute to technical architecture decisions and product strategy
Skill Required
- Strong programming skills with demonstrated experience in AI/ML frameworks (PyTorch, TensorFlow)
- Expertise in speech processing, Digital Signal Processing, and audio engineering
- Advanced Python programming
- Experience with Azure
- Proficiency in Real Time audio processing with target latency
- Experience optimising models for edge deployment
- Knowledge of audio compression techniques and format
- Familiarity with audio quality metrics
- Experience with audio processing libraries
- Proficiency in version control (Git) and CI/CD pipelines
- Previous work on TTS systems (commercial or lab)
- Background in voice conversion or voice cloning technologies
AI/ML Platform Experience
- Experience with Groq for high-performance inference
- Familiarity with Deepgram's API and speech-to-text capabilities
- Knowledge of large language model deployment and optimisation
Speech Technology Expertise
- Deep understanding of modern TTS architectures:
- Non-autoregressive models (FastSpeech 2, Glow-TTS)
- Autoregressive models (Tacotron 2, YourTTS)
- Flow-based models (Flow-TTS, WaveFlow)
- Experience with vocoders:
- HiFi-GAN
- WaveNet
- UnivNet
- BigVGAN
Share this job
Useful Links