{"jobs":[{"absolute_url":"https://job-boards.greenhouse.io/nuancelabs/jobs/4272077009","data_compliance":[{"type":"gdpr","requires_consent":false,"requires_processing_consent":false,"requires_retention_consent":false,"retention_period":null,"demographic_data_consent_applies":false}],"internal_job_id":4158658009,"location":{"name":"Seattle, Washington"},"metadata":null,"id":4272077009,"updated_at":"2026-06-05T18:25:27-04:00","requisition_id":"12","title":"Administrative Business Partner","company_name":"Nuance Labs","first_published":"2026-06-02T19:11:42-04:00","language":"en","application_deadline":null,"content":"\u0026lt;div class=\u0026quot;content-intro\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;About Nuance Labs\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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 real person.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re a Series A company ($60M raised) backed by Lightspeed, Accel, South Park Commons, NVentures, and Define Ventures, with PhDs from MIT, UW, Oxford, CMU, and Johns Hopkins, and industry experience from Apple, Meta, Amazon AGI, and Discord. The team is small, the work is real, and the problems are unsolved.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;How Nuance Differentiates\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s a 10x improvement, and it demands rethinking the entire stack.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s an extremely hard problem, and we\u0026#39;re developing foundation models designed for it from the ground up.\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;\u0026lt;div\u0026gt;\n\u0026lt;h2\u0026gt;\u0026lt;strong\u0026gt;About the Role\u0026lt;/strong\u0026gt;\u0026lt;/h2\u0026gt;\n\u0026lt;div\u0026gt;We\u0026#39;re looking for an experienced administrative business partner to shape our operational core, from office management and team culture to the systems that scale with us.\u0026lt;/div\u0026gt;\n\u0026lt;strong\u0026gt;\u0026lt;br\u0026gt;\u0026lt;/strong\u0026gt;\u0026lt;/div\u0026gt;\n\u0026lt;h2\u0026gt;\u0026lt;strong\u0026gt;What you\u0026#39;ll own\u0026lt;/strong\u0026gt;\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Our physical environment: own relationships with building and facilities management, and ensuring the space evolves with us as we grow. You\u0026#39;ll decide what a great in-person experience looks like here and make it happen\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Our expense management muscle: support expense tracking, own month-end close with our external accountant, and proactively find ways to make our spend smarter. You\u0026#39;ll have real visibility into the company\u0026#39;s financial operations from day one\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;How information flows: design and maintain the internal communications systems that keep a fast-moving team aligned\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;How we bring in people: build an on-site interview experience and onboarding experience that sets up every candidate and every new-hire to perform their best\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;How we come together: design the moments (e.g., team lunches, offsites, celebrations) that define our culture. At this stage, you have real influence over what kind of company we become\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Who we work with: manage our vendor and contract landscape, and proactively identify new partners and tools that unlock the next stage of growth. You\u0026#39;re not just maintaining relationships--you\u0026#39;re building them\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;\u0026lt;strong\u0026gt;Illustrative projects\u0026lt;/strong\u0026gt;\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Office redesign: partner with a designer and contractor team to reimagine our space for collaboration and productivity in an aesthetic that conveys our brand\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;IT support procurement: identify IT vendors that can help us scale our operations\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Annual team holiday party: own the full arc of our year-end celebration, from venue selection and vendor coordination to the agenda and details that make it feel like\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Compensation\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;$70,000 – $90,000 base salary.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;div class=\u0026quot;content-conclusion\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;Logistics\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Location:\u0026lt;/strong\u0026gt; In-person in Seattle, 5 days a week — we believe in the compounding value of working shoulder-to-shoulder\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Health:\u0026lt;/strong\u0026gt; HSA plan with ~$2,000 in company contributions — about 2x what most big tech companies offer\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;PTO:\u0026lt;/strong\u0026gt; 15 days + public holidays, and we close for a full week over the holidays\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Lunch, beverages, and snacks:\u0026lt;/strong\u0026gt; On us, every workday — the kind of thing that makes you actually look forward to the workday\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Commuter benefits\u0026lt;/strong\u0026gt;\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;401K:\u0026lt;/strong\u0026gt; In the works\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h3\u0026gt;\u0026lt;/h3\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;em\u0026gt;Nuance Labs is an equal opportunity employer. We believe diverse teams build better AI.\u0026lt;/em\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;","departments":[{"id":4031246009,"name":"Operations","child_ids":[],"parent_id":null}],"offices":[{"id":4030799009,"name":"Seattle","location":null,"child_ids":[],"parent_id":null}]},{"absolute_url":"https://job-boards.greenhouse.io/nuancelabs/jobs/4277601009","data_compliance":[{"type":"gdpr","requires_consent":false,"requires_processing_consent":false,"requires_retention_consent":false,"retention_period":null,"demographic_data_consent_applies":false}],"internal_job_id":4162947009,"location":{"name":"Seattle, Washington"},"metadata":null,"id":4277601009,"updated_at":"2026-06-05T18:20:40-04:00","requisition_id":"16","title":"Member of Technical Staff — ML Data Infra","company_name":"Nuance Labs","first_published":"2026-06-05T17:40:39-04:00","language":"en","application_deadline":null,"content":"\u0026lt;div class=\u0026quot;content-intro\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;About Nuance Labs\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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 real person.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re a Series A company ($60M raised) backed by Lightspeed, Accel, South Park Commons, NVentures, and Define Ventures, with PhDs from MIT, UW, Oxford, CMU, and Johns Hopkins, and industry experience from Apple, Meta, Amazon AGI, and Discord. The team is small, the work is real, and the problems are unsolved.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;How Nuance Differentiates\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s a 10x improvement, and it demands rethinking the entire stack.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s an extremely hard problem, and we\u0026#39;re developing foundation models designed for it from the ground up.\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;\u0026lt;h2\u0026gt;About the Role\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;Model quality is ultimately a data problem. The best architecture and the best training run can\u0026#39;t outrun bad, slow, or poorly curated data — and at the scale we\u0026#39;re operating, the difference between a good data pipeline and a great one shows up directly in the model.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re looking for someone who lives and breathes data at scale. You know how to build pipelines that are fast, reliable, and maintainable — and you\u0026#39;re just as comfortable taking a researcher\u0026#39;s messy processing script and turning it into something that runs on petabytes as you are designing a new pipeline architecture from scratch. Research moves fast here, and the ability to productionize quickly without losing fidelity is the core skill.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;Our data is multimodal — video, audio, and text — and the processing requirements are demanding: high throughput, low error rates, and strict quality filters. There\u0026#39;s a lot of interesting engineering work here, and the impact is direct and measurable.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;What You\u0026#39;ll Do\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Design, build, and operate large-scale data pipelines for ingestion, processing, filtering, and curation of multimodal training data (video, audio, text)\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Take research-grade data processing code and turn it into robust, production-level pipelines — quickly and without losing correctness\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Optimize pipeline throughput and efficiency at scale; identify and eliminate bottlenecks across compute, I/O, and storage\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Build and maintain data quality systems — deduplication, filtering, validation, and quality scoring at scale\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Manage petabyte-scale datasets: storage architecture, versioning, lineage tracking, and cost efficiency\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Work closely with researchers to understand data requirements and translate them into scalable processing systems\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Build tooling and infrastructure that makes the research team faster — efficient data access, reproducible processing, and fast iteration loops\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;What We\u0026#39;re Looking For\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Proven experience building and operating large-scale data pipelines in production — you\u0026#39;ve processed data at a scale where naive approaches break\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Strong proficiency with distributed data processing frameworks — Spark, Ray, Dask, or similar — and a clear sense of when to use each\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Solid software engineering fundamentals: you write clean, testable, maintainable code and understand why that matters when pipelines run unattended at scale\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with multimodal data (video, audio) is a strong plus — understanding of formats, codecs, and processing libraries (FFmpeg, decord, etc.)\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Familiarity with ML data pipelines specifically — understanding of how data quality and format affect model training\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Ability to move fast: you can take a prototype script from a researcher and ship a production version in days, not weeks\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Bonus Points\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Experience building data pipelines for large-scale model training (pre-training or fine-tuning)\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Familiarity with data versioning and lineage tools (DVC, Delta Lake, Apache Iceberg, etc.)\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with streaming data pipelines or online data processing\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Prior work at an AI lab, video platform, or other data-intensive company\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Contributions to open-source data tooling\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Compensation\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;$200,000 – $300,000 base salary, plus meaningful equity. We think long-term ownership matters and structure equity accordingly.\u0026lt;/p\u0026gt;\u0026lt;div class=\u0026quot;content-conclusion\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;Logistics\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Location:\u0026lt;/strong\u0026gt; In-person in Seattle, 5 days a week — we believe in the compounding value of working shoulder-to-shoulder\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Health:\u0026lt;/strong\u0026gt; HSA plan with ~$2,000 in company contributions — about 2x what most big tech companies offer\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;PTO:\u0026lt;/strong\u0026gt; 15 days + public holidays, and we close for a full week over the holidays\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Lunch, beverages, and snacks:\u0026lt;/strong\u0026gt; On us, every workday — the kind of thing that makes you actually look forward to the workday\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Commuter benefits\u0026lt;/strong\u0026gt;\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;401K:\u0026lt;/strong\u0026gt; In the works\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h3\u0026gt;\u0026lt;/h3\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;em\u0026gt;Nuance Labs is an equal opportunity employer. We believe diverse teams build better AI.\u0026lt;/em\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;","departments":[{"id":4031248009,"name":"Engineering","child_ids":[],"parent_id":null}],"offices":[{"id":4030799009,"name":"Seattle","location":null,"child_ids":[],"parent_id":null}]},{"absolute_url":"https://job-boards.greenhouse.io/nuancelabs/jobs/4277592009","data_compliance":[{"type":"gdpr","requires_consent":false,"requires_processing_consent":false,"requires_retention_consent":false,"retention_period":null,"demographic_data_consent_applies":false}],"internal_job_id":4162941009,"location":{"name":"Seattle, Washington"},"metadata":null,"id":4277592009,"updated_at":"2026-06-05T18:23:16-04:00","requisition_id":"15","title":"Member of Technical Staff — Model Optimization and Inference","company_name":"Nuance Labs","first_published":"2026-06-05T17:33:35-04:00","language":"en","application_deadline":null,"content":"\u0026lt;div class=\u0026quot;content-intro\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;About Nuance Labs\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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 real person.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re a Series A company ($60M raised) backed by Lightspeed, Accel, South Park Commons, NVentures, and Define Ventures, with PhDs from MIT, UW, Oxford, CMU, and Johns Hopkins, and industry experience from Apple, Meta, Amazon AGI, and Discord. The team is small, the work is real, and the problems are unsolved.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;How Nuance Differentiates\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s a 10x improvement, and it demands rethinking the entire stack.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s an extremely hard problem, and we\u0026#39;re developing foundation models designed for it from the ground up.\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;\u0026lt;h2\u0026gt;About the Role\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;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\u0026#39;ve worked with vLLM, SGLang, or similar frameworks and have opinions about where they fall short.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;Our stack is more complex than a standard LLM deployment: we\u0026#39;re serving a full-duplex multimodal system that must satisfy strict real-time latency constraints. There\u0026#39;s a lot of unsolved optimization work here, and we need someone who finds that genuinely exciting.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;What You\u0026#39;ll Do\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Own end-to-end inference optimization across our model stack — LLMs, audio models, and diffusion-based components\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Implement and tune KV cache strategies for long-context conversations, including eviction policies, compression, and memory-efficient attention\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Evaluate, deploy, and extend inference serving frameworks (vLLM, SGLang, TensorRT-LLM, etc.) for our specific workloads\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Profile and benchmark end-to-end latency and throughput; identify and systematically eliminate bottlenecks\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Build internal tooling that makes optimization work faster and more rigorous — profiling viewers, end-to-end inference test harnesses, and other infrastructure that helps the team move quickly\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Accelerate diffusion model inference — consistency models, step distillation, caching strategies, and custom kernel optimizations\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Apply and develop quantization techniques (INT8, INT4, GPTQ, AWQ, and beyond) to reduce memory footprint and increase throughput without meaningfully degrading quality\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Work closely with research and infrastructure to ensure new models ship with optimized serving from day one\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;What We\u0026#39;re Looking For\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Deep expertise in LLM inference optimization — you\u0026#39;ve worked on KV caching, memory layout, attention kernels, or batching strategies in a production or research context\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;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\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience optimizing diffusion model inference (latency reduction, step distillation, caching, or kernel-level work)\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Strong Python and PyTorch skills; comfort reading and writing CUDA or Triton kernels is a significant plus\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;A systematic approach to profiling and optimization — you measure first, then optimize\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Familiarity with speculative decoding or other inference-time acceleration techniques\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Bonus Points\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Hands-on experience with post-training quantization (GPTQ, AWQ, or similar) and understanding of quality/performance tradeoffs\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Familiarity with multimodal or streaming inference architectures\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience deploying real-time AI systems with hard latency SLAs\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Prior work at an AI lab, inference startup, or on a high-traffic model serving platform\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Contributions to open-source inference frameworks\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Compensation\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;$250,000 – $350,000 base salary, plus meaningful equity. We think long-term ownership matters and structure equity accordingly.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;div class=\u0026quot;content-conclusion\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;Logistics\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Location:\u0026lt;/strong\u0026gt; In-person in Seattle, 5 days a week — we believe in the compounding value of working shoulder-to-shoulder\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Health:\u0026lt;/strong\u0026gt; HSA plan with ~$2,000 in company contributions — about 2x what most big tech companies offer\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;PTO:\u0026lt;/strong\u0026gt; 15 days + public holidays, and we close for a full week over the holidays\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Lunch, beverages, and snacks:\u0026lt;/strong\u0026gt; On us, every workday — the kind of thing that makes you actually look forward to the workday\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Commuter benefits\u0026lt;/strong\u0026gt;\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;401K:\u0026lt;/strong\u0026gt; In the works\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h3\u0026gt;\u0026lt;/h3\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;em\u0026gt;Nuance Labs is an equal opportunity employer. We believe diverse teams build better AI.\u0026lt;/em\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;","departments":[{"id":4031247009,"name":"Research","child_ids":[],"parent_id":null}],"offices":[{"id":4030799009,"name":"Seattle","location":null,"child_ids":[],"parent_id":null}]},{"absolute_url":"https://job-boards.greenhouse.io/nuancelabs/jobs/4274385009","data_compliance":[{"type":"gdpr","requires_consent":false,"requires_processing_consent":false,"requires_retention_consent":false,"retention_period":null,"demographic_data_consent_applies":false}],"internal_job_id":4160141009,"location":{"name":"Seattle, Washington"},"metadata":null,"id":4274385009,"updated_at":"2026-06-05T17:08:48-04:00","requisition_id":"13","title":"Member of Technical Staff — Pretraining Infra","company_name":"Nuance Labs","first_published":"2026-06-04T21:37:45-04:00","language":"en","application_deadline":null,"content":"\u0026lt;div class=\u0026quot;content-intro\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;About Nuance Labs\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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 real person.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re a Series A company ($60M raised) backed by Lightspeed, Accel, South Park Commons, NVentures, and Define Ventures, with PhDs from MIT, UW, Oxford, CMU, and Johns Hopkins, and industry experience from Apple, Meta, Amazon AGI, and Discord. The team is small, the work is real, and the problems are unsolved.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;How Nuance Differentiates\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s a 10x improvement, and it demands rethinking the entire stack.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s an extremely hard problem, and we\u0026#39;re developing foundation models designed for it from the ground up.\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;\u0026lt;h2\u0026gt;About the Role\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re looking for a deeply technical MTS to own distributed training infrastructure for large-scale omni model pretraining.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;This role sits at the intersection of research, systems, and GPU-scale execution — building the training stack from 0→1 and scaling it: distributed execution, parallelism, GPU communication, data loading, checkpointing, observability, and debugging.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;Our models are omni from the ground up (audio, video, language, real-time full-duplex), which introduces systems challenges beyond standard LLM training: multimodal synchronization, long temporal context, variable sequence lengths, and tight memory/throughput constraints.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;High ownership. Direct impact on what models we can train, how fast research can iterate, and how reliably we scale.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;What You’ll Own\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Own the distributed training stack for omni model pretraining, from 0→1 system design to 1→10 scaling across large GPU clusters.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Build and operate the core training runtime: job orchestration, distributed execution, checkpointing, recovery, monitoring, and debugging for long-running training jobs.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Optimize large-scale training performance across parallelism strategy, GPU communication, memory usage, data throughput, MFU, step time, and end-to-end training efficiency.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Build infrastructure for omni training workloads: high-throughput audio/video/text data loading, temporal alignment, variable sequence handling, multimodal synchronization, and memory-efficient training.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Evolve the platform as model architectures, training recipes, data mixtures, sequence lengths, hardware constraints, and research directions change.\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;What We’re Looking For\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Hands-on experience running large-scale distributed training jobs across large GPU clusters; experience at hundreds of GPUs minimum, 1,000+ GPUs a strong plus.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Deep understanding of distributed training mechanics: data/tensor/pipeline/sequence parallelism, gradient communication, collectives, mixed precision, activation checkpointing, optimizer state, memory pressure, and framework-level tradeoffs.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Strong understanding of GPU communication and performance debugging: NCCL, all-reduce/all-gather/reduce-scatter, communication-computation overlap, topology, synchronization, stragglers, low MFU, OOMs, checkpoint bottlenecks, and data starvation.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Practical experience with at least one major large-scale training stack such as Megatron, PyTorch FSDP, DeepSpeed, or equivalent internal infrastructure.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Understanding of omni or multimodal training challenges, especially audio/video/language data, long temporal context, variable sequence lengths, modality-specific bottlenecks, and high-throughput dataloading.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Strong software engineering fundamentals, curiosity, and adaptability to new model architectures, training frameworks, hardware constraints, and research ideas.\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Bonus Points\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Prior 0→1 experience building large-scale training infrastructure or deeply modifying core training frameworks, runtimes, checkpointing, or debugging systems.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience training large omni or multimodal models involving audio, video, text, or long-context temporal data.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with adjacent infrastructure areas such as RL/post-training, data infrastructure, synthetic data generation, evaluation, or serving.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Publications or substantial open-source contributions in ML systems, distributed systems, HPC, GPU performance, or training infrastructure.\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Compensation\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;$300,000 – $400,000 base salary, plus meaningful equity. We think long-term ownership matters and structure equity accordingly.\u0026lt;/p\u0026gt;\u0026lt;div class=\u0026quot;content-conclusion\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;Logistics\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Location:\u0026lt;/strong\u0026gt; In-person in Seattle, 5 days a week — we believe in the compounding value of working shoulder-to-shoulder\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Health:\u0026lt;/strong\u0026gt; HSA plan with ~$2,000 in company contributions — about 2x what most big tech companies offer\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;PTO:\u0026lt;/strong\u0026gt; 15 days + public holidays, and we close for a full week over the holidays\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Lunch, beverages, and snacks:\u0026lt;/strong\u0026gt; On us, every workday — the kind of thing that makes you actually look forward to the workday\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Commuter benefits\u0026lt;/strong\u0026gt;\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;401K:\u0026lt;/strong\u0026gt; In the works\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h3\u0026gt;\u0026lt;/h3\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;em\u0026gt;Nuance Labs is an equal opportunity employer. We believe diverse teams build better AI.\u0026lt;/em\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;","departments":[{"id":4031248009,"name":"Engineering","child_ids":[],"parent_id":null}],"offices":[{"id":4030799009,"name":"Seattle","location":null,"child_ids":[],"parent_id":null}]},{"absolute_url":"https://job-boards.greenhouse.io/nuancelabs/jobs/4277561009","data_compliance":[{"type":"gdpr","requires_consent":false,"requires_processing_consent":false,"requires_retention_consent":false,"retention_period":null,"demographic_data_consent_applies":false}],"internal_job_id":4162923009,"location":{"name":"Seattle, Washington"},"metadata":null,"id":4277561009,"updated_at":"2026-06-05T17:13:11-04:00","requisition_id":"14","title":"Member of Technical Staff — RL Research","company_name":"Nuance Labs","first_published":"2026-06-05T17:13:11-04:00","language":"en","application_deadline":null,"content":"\u0026lt;div class=\u0026quot;content-intro\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;About Nuance Labs\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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 real person.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;We\u0026#39;re a Series A company ($60M raised) backed by Lightspeed, Accel, South Park Commons, NVentures, and Define Ventures, with PhDs from MIT, UW, Oxford, CMU, and Johns Hopkins, and industry experience from Apple, Meta, Amazon AGI, and Discord. The team is small, the work is real, and the problems are unsolved.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;How Nuance Differentiates\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s a 10x improvement, and it demands rethinking the entire stack.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;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\u0026#39;s an extremely hard problem, and we\u0026#39;re developing foundation models designed for it from the ground up.\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;\u0026lt;h2\u0026gt;About the Role\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;We’re looking for a deeply technical Member of Technical Staff to own RL and post-training for large-scale omni models.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;This role is broader than a traditional RL algorithm role. You will be expected to understand modern post-training methods and build the infrastructure needed to run them at scale. The work spans RL method development, rollout generation, reward modeling, policy optimization, evaluation, data feedback loops, serving, observability, and distributed execution.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;You will build Nuance’s RL/post-training stack from 0→1 and scale it from 1→10. That means turning rapidly evolving research ideas into reliable training systems: defining the abstractions, choosing or modifying frameworks, wiring together rollout workers and trainers, building reward/evaluation loops, debugging failure modes, and making the system fast enough for researchers to iterate.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;For Nuance, post-training is not limited to text. Our models are omni from the ground up: audio, video, language, and real-time full-duplex interaction. We need RL and post-training methods that improve interactive behavior, timing, interruption, emotional response, audiovisual coherence, and real-time conversational quality.\u0026lt;/p\u0026gt;\n\u0026lt;p\u0026gt;This is a high-ownership role with direct impact on how Nuance models improve after pretraining.\u0026lt;/p\u0026gt;\n\u0026lt;h2\u0026gt;What You’ll Own\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Build Nuance’s RL/post-training stack from 0→1: rollout generation, policy optimization, reward/reference model serving, data feedback loops, evaluation, checkpointing, observability, and debugging.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Develop and scale post-training methods such as PPO, GRPO, DPO, rejection sampling, RLHF/RLAIF, online RL, and model-based data improvement.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Design the systems abstractions that connect research ideas to production-scale RL runs: trainers, rollout workers, reward models, evaluators, data queues, experience buffers, and checkpoint promotion.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Build evaluation and feedback loops for omni behavior: turn-taking, interruption, timing, emotional response, audiovisual coherence, instruction following, and real-time interaction quality.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Optimize the end-to-end post-training loop across rollout throughput, serving latency, GPU utilization, policy update efficiency, queueing, checkpoint overhead, and research iteration speed.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Evolve the platform as algorithms, model architectures, reward definitions, data sources, and evaluation methods change.\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;What We’re Looking For\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Hands-on experience with RL, RLHF, RLAIF, post-training, alignment, or large-scale fine-tuning for modern foundation models.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Strong understanding of RL/post-training methods: policy optimization, reward modeling, preference optimization, rejection sampling, KL control, evaluation, and data feedback loops.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Ability to reason about model behavior and training dynamics: reward hacking, unstable rewards, distribution shift, stale policies, mode collapse, over-optimization, noisy preferences, and evaluation mismatch.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Practical experience building or operating RL/post-training pipelines with frameworks such as verl, ms-swift, OpenRLHF, or equivalent internal systems, including integration with rollout serving systems such as vLLM.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with large-scale training or inference systems, including rollout generation, model serving, batching, queueing, GPU utilization, checkpointing, and debugging.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Understanding of omni post-training for real-time audio-video-language interaction: temporal alignment, interruption, emotional response, and multimodal evaluation.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Strong software engineering fundamentals, curiosity, and adaptability to new RL algorithms, model architectures, serving systems, evaluation methods, and research ideas.\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Bonus Points\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;Prior 0→1 experience building post-training systems, RL pipelines, agent training systems, evaluation platforms, or large-scale model improvement loops.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with PPO, GRPO, DPO, online RL, RLHF/RLAIF, reward modeling, preference data, synthetic data generation, or model-based data improvement.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with omni or multimodal post-training for audio-video-language models, especially long-context or real-time interactive systems.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience scaling mixed training/inference workloads across large GPU clusters.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Experience with adjacent areas such as distributed pretraining, data infrastructure, inference serving, simulation, human/AI feedback collection, or evaluation infrastructure.\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;Publications or substantial open-source contributions in RL, post-training, alignment, evaluation, ML systems, or model behavior.\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h2\u0026gt;Compensation\u0026lt;/h2\u0026gt;\n\u0026lt;p\u0026gt;$300,000 – $400,000 base salary, plus meaningful equity. We think long-term ownership matters and structure equity accordingly.\u0026lt;/p\u0026gt;\u0026lt;div class=\u0026quot;content-conclusion\u0026quot;\u0026gt;\u0026lt;h2\u0026gt;Logistics\u0026lt;/h2\u0026gt;\n\u0026lt;ul\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Location:\u0026lt;/strong\u0026gt; In-person in Seattle, 5 days a week — we believe in the compounding value of working shoulder-to-shoulder\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Health:\u0026lt;/strong\u0026gt; HSA plan with ~$2,000 in company contributions — about 2x what most big tech companies offer\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;PTO:\u0026lt;/strong\u0026gt; 15 days + public holidays, and we close for a full week over the holidays\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Lunch, beverages, and snacks:\u0026lt;/strong\u0026gt; On us, every workday — the kind of thing that makes you actually look forward to the workday\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;Commuter benefits\u0026lt;/strong\u0026gt;\u0026lt;/li\u0026gt;\n\u0026lt;li\u0026gt;\u0026lt;strong\u0026gt;401K:\u0026lt;/strong\u0026gt; In the works\u0026lt;/li\u0026gt;\n\u0026lt;/ul\u0026gt;\n\u0026lt;h3\u0026gt;\u0026lt;/h3\u0026gt;\n\u0026lt;p\u0026gt;\u0026lt;em\u0026gt;Nuance Labs is an equal opportunity employer. We believe diverse teams build better AI.\u0026lt;/em\u0026gt;\u0026lt;/p\u0026gt;\u0026lt;/div\u0026gt;","departments":[{"id":4031247009,"name":"Research","child_ids":[],"parent_id":null}],"offices":[{"id":4030799009,"name":"Seattle","location":null,"child_ids":[],"parent_id":null}]}],"meta":{"total":5}}