Member of Technical Staff — Model Optimization and Inference

Nuance Labs•Seattle, WA

8d•$200,000 - $400,000•Onsite

About The Position

Nuance Labs is building photorealistic, real-time AI avatars with emotional intelligence: a full-duplex audiovisual system that can listen, speak, react, interrupt, and respond like a human. The team is small, the work is real, and the problems are unsolved. Most conversational AI avatars today are hacks — a face slapped on a speech-to-speech pipeline, stuck in the uncanny valley: emotionless, mechanical, one-turn-at-a-time. Current systems take 2–5 seconds to respond; natural conversation requires sub-500ms. That's a 10x improvement, and it demands rethinking the entire stack. That rethinking starts with full-duplex: an AI that listens and speaks simultaneously, perceives emotion in real time, and responds with a face that actually reflects it. It's an extremely hard problem, and we're developing foundation models designed for it from the ground up. We can train a great model. The next problem is making it fast enough to actually use in a real-time conversation — and that gap is enormous. A model that responds in 3 seconds is a demo. A model that responds in under 500ms is a product. We're looking for someone who specializes in taking trained models and squeezing every last millisecond out of them. You understand the full stack from model weights to serving infrastructure — quantization, KV cache optimization, kernel-level acceleration, batching strategies — and you know which lever to pull for which problem. You've worked with vLLM, SGLang, or similar frameworks and have opinions about where they fall short. Our stack is more complex than a standard LLM deployment: we're serving a full-duplex multimodal system that must satisfy strict real-time latency constraints. There's a lot of unsolved optimization work here, and we need someone who finds that genuinely exciting.

Requirements

Deep expertise in LLM inference optimization — you've worked on KV caching, memory layout, attention kernels, or batching strategies in a production or research context
Hands-on experience with quantization for LLMs (post-training quantization, QAT, or both) and understanding of quality/performance tradeoffs
Proficiency with inference serving frameworks — vLLM, SGLang, TensorRT-LLM, or similar — including the ability to go beyond default configurations and adapt them to non-standard use cases
Experience optimizing diffusion model inference (latency reduction, step distillation, caching, or kernel-level work)
Strong Python and PyTorch skills; comfort reading and writing CUDA or Triton kernels is a significant plus
A systematic approach to profiling and optimization — you measure first, then optimize

Nice To Haves

Experience with speculative decoding or other inference-time acceleration techniques
Familiarity with multimodal or streaming inference architectures
Experience deploying real-time AI systems with hard latency SLAs
Prior work at an AI lab, inference startup, or on a high-traffic model serving platform
Contributions to open-source inference frameworks

Responsibilities

Own end-to-end inference optimization across our model stack — LLMs, audio models, and diffusion-based components
Implement and tune KV cache strategies for long-context conversations, including eviction policies, compression, and memory-efficient attention
Apply and develop quantization techniques (INT8, INT4, GPTQ, AWQ, and beyond) to reduce memory footprint and increase throughput without meaningfully degrading quality
Accelerate diffusion model inference — consistency models, step distillation, caching strategies, and custom kernel optimizations
Evaluate, deploy, and extend inference serving frameworks (vLLM, SGLang, TensorRT-LLM, etc.) for our specific workloads
Profile and benchmark end-to-end latency and throughput; identify and systematically eliminate bottlenecks
Work closely with research and infrastructure to ensure new models ship with optimized serving from day one