Research Engineer, Audio Models (Remote — Worldwide)

Join the world's leading AI audio research team to push the boundaries of speech synthesis, voice cloning, and generative media. As a Research Engineer, you will design and scale the next generation of foundation models that power ElevenLabs’ industry-leading API. Your work will directly impact millions of users by improving the latency, prosody, and emotional depth of real-time audio generation.

- Design and implement novel neural architectures for high-fidelity speech synthesis and audio generation. - Scale training pipelines to handle massive, multi-lingual datasets for foundation model pre-training. - Optimize inference engines to achieve sub-second latency for real-time conversational AI applications. - Experiment with new techniques in prompt engineering, fine-tuning, and zero-shot voice cloning. - Collaborate with the infrastructure team to deploy models on distributed H100/A100 clusters. - Stay current with SOTA research in TTS, vocoders, and audio-visual synchronization to maintain our competitive edge. - Conduct rigorous evaluations of model performance using both automated metrics and human-in-the-loop feedback.

- MS or PhD in Computer Science, Machine Learning, or a related field, or equivalent industry experience. - Strong publication record or extensive portfolio of projects involving Large Language Models (LLMs) or Audio Generation. - Deep expertise in PyTorch and training large-scale models across distributed GPU clusters. - Experience with modern generative architectures including Diffusion, Transformers, and VAEs. - Proficiency in C++ and CUDA for optimizing model kernels and inference speed. - Proven ability to bridge the gap between academic research and production-ready code. - Self-directed mindset capable of thriving in a fast-paced, remote-first environment across time zones.

Equity, benefits, remote.