Machine Learning Engineer, Speech - Joint Audio-Video Modeling

We're looking for a Research / ML Engineer to join our Speech Team to build state-of-the-art speech and audio generation systems end-to-end from data specs through production inference with a focus on joint audio-video modeling.

You'll own the audio side of multimodal generation: the representations (audio VAEs, neural codecs), the generative backbone (diffusion / flow-matching transformers), and the conditioning and alignment machinery that makes characters speak, sing, and emote in sync with what's on screen. That includes voice cloning and multi-speaker conditioning inside joint AV models, cinematic dialogue with music and sound design, and adjacent speech tasks (controllable TTS, voice conversion) that feed the same stack.

You'll drive the model ↔ data ↔ eval flywheel, partnering closely with research, video, data, and infra to ship fast, reliable, and cost-aware models. In this role you'll work at the intersection of cutting-edge research and practical engineering, contributing to the development of safe, steerable, and trustworthy AI systems.

You will thrive in this role if you:

See research and engineering as two sides of the same coin and enjoy owning work end-to-end.

Are excited to work across modalities and collaborate closely with a video generation team rather than staying inside audio.

Are results-oriented, flexible, and willing to pick up whatever moves the needle.

Like collaborating closely with infra, data, and product to ship measurable improvements.

Enjoy designing experiments, listening tests, and metrics that correlate with user-perceived quality.

Are eager to learn every day, and to find and solve unique large-scale problems.

Audio Representations: Design, train, and improve the audio VAEs, neural codecs, and vocoders our generative models sit on top of latent design, reconstruction and perceptual objectives, compression-vs-fidelity tradeoffs.

Model Building: Architect, implement, pre-train, fine-tune, and post-train/alignment (e.g., GRPO/DPO) diffusion and flow-matching transformers for large-scale audio and video generation.

Joint Audio-Video Modeling: Design the audio conditioning and cross-modal alignment inside joint AV models, audio latents alongside video latents, reference-audio and multi-speaker conditioning, multi shot generation audio/video modeling.

Experimental Design: Design, run, and analyze scientific experiments to advance our understanding of the models.

Data Ownership: Define data requirements and collaborate on acquisition, curation, AV-sync and quality filtering, annotation quality, and synthetic data strategies for paired audio-video and speech corpora.

Rigorous Evaluation: Design automated objective/subjective evaluations audio fidelity and intelligibility metrics, AV-sync, listening and viewing tests, robustness & bias checks, and red-team studies.

Inference Efficiency: Drive distillation, step-count reduction, quantization, and kernel/memory optimization to meet interactive latency and cost targets.

Pipeline Delivery: Harden the training → evaluation → inference pipeline; profile latency, memory, and cost; and meet production SLAs with robust monitoring and rollback.

GPU Scaling: Partner with infrastructure to run distributed training/inference on cloud fleets and productionize models with reliability and observability.

Project Leadership: Independently lead small research projects while collaborating on larger team initiatives, including cross-team work with video generation.

Tool Development: Develop and improve dev tooling to enhance team productivity.

Safety & Responsibility: Contribute to safety/consent guardrails, watermarking, and misuse/abuse mitigation for responsible voice and likeness technology.

Exceptional research/development experience with large-scale audio models (>8B parameters, >500k hours of data).

Deep hands-on experience with diffusion and/or flow-matching transformers, including practical knowledge of samplers, schedules, conditioning mechanisms, and distillation.

Deep hands-on experience training audio VAEs, neural audio codecs, and vocoders latent/tokenizer design, reconstruction and perceptual objectives, adversarial training.

Strong experience with multi-node, multi-GPU distributed training (FSDP/DeepSpeed or equivalent).

Strong software engineering skills with a proven track record of building complex systems.

Strong with PyTorch and performance work (profiling, CUDA/Triton/C++ as needed) and writing reliable production-quality code.

Shipped large-scale speech/audio or multimodal generative models to production.

Background in working with large-scale ML data, and the ability to iterate on data and triangulate quality using both subjective and objective signals.

Experience with voice cloning, speech control/steerability, or expressive speech generation.

Notable publications and/or open-source contributions in speech/audio/ML.

Strongly preferred: Experience with multimodal audio-video modeling : joint AV generation of multi-shot, multi-speaker scenes with dialogue, music, and sound design generated jointly with video, and the cross-modal alignment that keeps them in sync. Experience with video generation : video diffusion/flow-matching transformers, video VAEs, conditioned and multi-shot generation, building data pipelines for video models. Streaming or real-time generation, causal distillation (e.g., Self Forcing / Self Forcing++).

Experience with multimodal audio-video modeling : joint AV generation of multi-shot, multi-speaker scenes with dialogue, music, and sound design generated jointly with video, and the cross-modal alignment that keeps them in sync.

Experience with video generation : video diffusion/flow-matching transformers, video VAEs, conditioned and multi-shot generation, building data pipelines for video models.

Streaming or real-time generation, causal distillation (e.g., Self Forcing / Self Forcing++).