What are the responsibilities and job description for the Machine Learning Infra Engineer position at Nuance Labs?
About The Role
Nuance Labs is building the next generation of emotionally expressive, real-time AI.
This is a critical role to build the infrastructure that powers our AI platform. You will own the systems that serve models at scale, orchestrate complex data workflows, and ensure our real-time video AI runs reliably with low latency for users worldwide.
Responsibilities
Nuance Labs is building the next generation of emotionally expressive, real-time AI.
This is a critical role to build the infrastructure that powers our AI platform. You will own the systems that serve models at scale, orchestrate complex data workflows, and ensure our real-time video AI runs reliably with low latency for users worldwide.
Responsibilities
- Own Inference Infrastructure: Build and maintain the serving stack for multimodal AI workloads. Optimize for latency, throughput, and cost using batching strategies, autoscaling, and intelligent resource allocation.
- Real-Time Video Streaming: Architect systems to handle long-lived WebRTC connections with unpredictable client behavior, ensuring smooth video and audio delivery at scale.
- Orchestrate Data Workflows: Build robust pipelines for offline processing, evaluation, and training using orchestration frameworks like Dagster or Ray. Manage petabyte-scale video storage and network requirements.
- GPU Cluster Management: Configure and maintain GPU clusters using Kubernetes and Terraform. Implement monitoring, autoscaling based on custom metrics, and cost optimization strategies.
- Developer Tooling: Build CI/CD, evaluation, and versioning systems that enable safe, zero-downtime model deployments and rapid iteration cycles.
- Infrastructure Expertise: Strong practical experience with Kubernetes, Terraform, and cloud platforms. You can design secure, scalable systems and debug complex distributed issues.
- Systems Programming: Proficiency in Python and experience with systems languages (Rust or Go). Comfortable profiling workloads and resolving compute, memory, or network bottlenecks.
- Orchestration & Pipelines: Experience managing large-scale offline workflows using tools like Dagster, Ray, Airflow, or similar frameworks.
- Production Operations: Deep understanding of production reliability, monitoring, incident response, and capacity planning for high-traffic services.
- Experience with WebRTC or real-time media pipelines in production
- Experience running GPU-backed inference services at scale (vLLM, Triton Inference Server, TensorRT)
- Knowledge of performance optimization and low-level systems debugging
- Familiarity with video/audio processing and storage systems
- $10M seed round backed by Accel, South Park Commons, Lightspeed, and top angels including Synthesia’s former CPO.
- A world-class team of PhDs from MIT, UW, and Oxford with decades of industry experience at Apple and Meta, advancing real-time avatars from cutting-edge research to products used by millions.
- In-person collaboration, 5 days a week at Seattle HQ