Sr Staff R&D Engineer

What You’ll Do

• Lead the research, design, and implementation of state-of-the-art machine learning algorithms for speech processing, voice transfer, source separation, and upmixing in media post-production environments.
• Drive the architecture and deployment of scalable model training pipelines using PyTorch and distributed computing frameworks.
• Develop novel generative audio models, including latent diffusion, flow-based models, variational autoencoders, and neural vocoders, optimized for professional soundtrack production.
• Own end-to-end model lifecycle management: pretraining, fine-tuning, validation, inference optimization, and CI/CD integration.
• Guide the development of personalized model adaptation workflows to support per-user tuning, cross-project continuity, and flexible deployment.
• Collaborate with product, platform, and engineering leads to define integration strategies within a secure, cloud-optimized SaaS environment.
• Stay at the forefront of generative audio, multi-modal modeling, and self-supervised learning—translating emerging research into applied innovation.
• Contribute to internal tooling and infrastructure that improves iteration speed, reproducibility, and explainability of deployed models.
• Mentor junior researchers and engineers, and contribute to a culture of rigorous experimentation, collaboration, and continuous improvement.

What We’re Looking For

• MSc or PhD in Computer Science, Electrical Engineering, Applied Math, or a related field with a focus on AI/ML and mult-imodal signal processing.
• 5 years of professional experience in applied ML, with a deep focus on audio-centric AI/ML research and deployment.
• Expertise in building and scaling models using PyTorch, with fluency in training, fine-tuning, and inference for deep neural networks.
• Demonstrated experience developing generative models such as VAE, GAN, diffusion models, or neural vocoders (e.g., HiFi-GAN, WaveNet).
• Deep understanding of audio-specific ML domains, including source separation, speech enhancement, music processing, and cross-modal tasks.
• Experience with MLOps tooling (e.g., Weights & Biases, MLflow, Datachain), Docker-based containerization, and scalable infrastructure for distributed training.
• Fluency in audio signal processing fundamentals and the integration of DSP into ML pipelines.
• Proven ability to contribute to architectural planning, research strategy, and production deployment in complex, multi-stakeholder environments.

Preferred Qualifications

• Familiarity with audio/text/video multi-modal frameworks and cross-domain representations.
• Experience implementing real-time or near-real-time inference pipelines in cloud or edge environments (e.g., AWS, GCP, on-prem GPUs).
• Working knowledge of latent diffusion audio models (e.g., stable-audio, AudioLDM, AudioGen).
• Strong knowledge of industry-standard audio datasets and benchmarks (LibriSpeech, VCTK, MUSDB, etc.).
• Experience optimizing inference pipelines for creative applications or interactive use.
• Proficiency in lower-level audio frameworks (C / C++, etc.)
• Contributions to published research at top-tier conferences (NeurIPS, ICASSP, ICLR, Interspeech) and/or open-source ML frameworks.