Research Staff – Voice AI Foundations

This is a fully remote position, open to applicants in Australia.

📋 Description

• Create cutting-edge neural audio codecs featuring ultra-low-bit-rate compression and high-fidelity reconstruction.

• Lead the development of steerable generative models for a variety of human speech scenarios, including emotional tones, multiple speakers, background noise, and overlapping speech.

• Construct embedding systems that decompose latent representations into dimensions of speaker, content, style, environment, and channel.

• Utilize latent recombination techniques to produce synthetic audio data on scales previously deemed unfeasible.

• Train multimodal speech-to-speech systems to ensure robust comprehension and empathetic, human-like interactions.

• Design model architectures, training methodologies, and inference algorithms optimized for bare-metal hardware performance.

• Facilitate cost-effective training on datasets encompassing billions of hours, enabling real-time inference for millions of simultaneous conversations.

• Execute thorough controlled experiments, ablation studies, evaluations, stress testing, and analyses of edge cases.

• Contribute to pioneering research that advances the fields of voice AI, speech, and language modeling.

⛳️ Requirements

• Solid mathematical grounding in statistical learning theory, especially in the areas of self-supervised and multimodal learning.

• Profound knowledge of foundational model architectures and the ability to scale training across various modalities.

• Capacity to derive innovative mathematical formulations and implement them in an efficient manner.

• Proven experience in constructing and curating extensive datasets while ensuring quality and diversity.

• Demonstrated history of designing controlled experiments that validate architectural innovations and theoretical concepts.

• Experience in optimizing models for deployment in real-world settings, taking into account hardware limitations and efficiency strategies.

• Background in open-source contributions or research publications that advance the fields of speech and language AI.

• Comfort in actively using and experimenting with advanced AI tools.

• Ability to quickly identify critical experiments that can validate or refute ideas.

• Vision to scale successful proofs-of-concept by a factor of 100.

• Familiarity with neural audio codecs, generative models, embedding systems, latent representations, multimodal speech-to-speech systems, model architectures, training methodologies, inference algorithms, and efficient hardware deployment.

🏝️ Benefits

• Flexible remote work options.

• An AI-driven work environment that promotes the use and experimentation with advanced AI tools.

• An opportunity to engage in transformative voice AI research at scale.

• Opportunities for open-source contributions and research publications.

• The option to decline AI Notetaker interview recordings without impacting candidacy.