Nuance Labs

Designs and operates human-evaluation studies for real-time AI avatars, translating subjective qualities such as naturalness, emotion, trust, and presence into reliable signals for model development and release decisions. Requires 5+ years of qualitative and quantitative human-subjects research experience.

3-month research fellowship for early-career researchers working on frontier Multimodal LLMs, generative modeling, and real-time audiovisual AI. Own a research problem in pretraining, post-training, RL, evaluation, or multimodal modeling. Strong PyTorch and first-author tier-1 paper required.

New/recent PhD to own RL and post-training for large-scale omni models. Build and scale the full RL/post-training stack including rollout, optimization, reward modeling, and evaluation for real-time audiovisual AI.

Early-career engineer optimizing inference for real-time multimodal AI avatars. Focus on KV cache strategies, serving frameworks, quantization, and latency reduction for LLMs and diffusion models.

Optimize inference for real-time multimodal AI avatars. Specialize in LLM and diffusion model serving, KV cache strategies, quantization, and low-latency frameworks like vLLM and TensorRT-LLM.

Build and operate large-scale multimodal data pipelines for AI avatar model training. Design production-grade systems for petabyte-scale video, audio, and text data.

Own RL and post-training infrastructure for omni foundation models. Build and scale rollout, reward, and policy systems from 0→1 for real-time audiovisual AI.

Own and scale the distributed training infrastructure for large-scale omni model pretraining across GPU clusters, covering job orchestration, parallelism, GPU communication, data loading, and performance optimization.