Nuance Labs
Member of Technical Staff — Model Optimization and Inference (Experienced)
$250K–$350K
Seattle, Washington +1 · Staff · Full-time · Posted Jun 5, 2026
- Location
- Seattle, Washington
Seattle - Experience
- Experienced engineers and researchers who have operated at senior to senior-staff level; no specific number of years stated.
- Visa sponsorship
- Visa sponsorship offered from day one (O-1, H-1B, green card, etc.).
About the job
Optimize and deploy low-latency inference for Nuance Labs’ real-time, full-duplex multimodal AI avatar system.
What you'll do
- Optimize inference end to end across LLMs, audio models, and diffusion components, including KV caching, quantization, and acceleration.
- Evaluate, deploy, and extend serving frameworks; profile latency and throughput and remove bottlenecks.
- Build optimization tooling and collaborate with research and infrastructure teams to ship models with efficient serving.
What they ask for
- Hands-on experience optimizing LLM inference in production or a high-traffic research context.
- Proficiency with vLLM, SGLang, TensorRT-LLM, or similar inference serving frameworks, including adapting them to non-standard workloads.
- Experience optimizing diffusion model inference.
- Strong Python and PyTorch skills.
Nice to have
- Comfort reading and writing CUDA or Triton kernels; familiarity with speculative decoding or other inference-time acceleration.
- Experience with post-training quantization, multimodal or streaming inference, real-time AI systems with hard latency SLAs, or open-source inference frameworks.
Pay
$250,000 – $350,000 USD base salary
Tools
Python, PyTorch, CUDA, Triton, vLLM, SGLang, TensorRT-LLM, GPTQ
AI focus
Optimize inference for real-time, full-duplex multimodal AI, including LLM, audio, and diffusion models, to meet strict latency constraints.
Model-written summary · the original posting is authoritative