Software Engineer, ML Serving

Rime Labs•San Francisco Bay Area, CA

18h•Onsite

About The Position

Rime is a foundation modeling company that builds voice AI for enterprises running customer experiences at scale. Our models are purpose-built for high-volume conversational deployments, engineered for the accuracy, performance, and deployment flexibility that production environments actually demand. We started from a different premise than the rest of the field: build voice AI for human connection, not slop. Before we trained a single model, we built our own corpus: full-duplex, studio-quality conversational speech of normal people, recorded and annotated by linguists. It's why our models are unparalleled in naturalism, and it's why enterprises pick Rime when pilots need to make it to production. We're hiring a Software Engineer to own the serving infrastructure that connects Rime's inference engines to the world. This role sits at the intersection of ML systems and cloud infrastructure — you'll work directly on model inference and cloud infrastructure to build, harden, and scale the systems that stream voice at real-time latency. As Rime moves toward its next-generation architecture, you'll be a core architect of how our models get served.

Requirements

Hands-on experience with real-time multinode ML serving infrastructure — ML serving framework experience: NVIDIA Dynamo/Triton, vLLM, SGLang, or equivalent.
Experience with distributed or disaggregated model serving (Tensor Parallel, Pipeline Parallel, or equivalent).
Strong cloud infrastructure fundamentals: Linux internals, networking, containerization (Docker, Kubernetes).
IaC experience — Terraform, Packer, or comparable.
You should have opinions about how to do this right.
On-call is part of the job.
You treat production reliability as a shared responsibility.

Nice To Haves

Experience with multinode training (DDP, FSDP, etc.).
Experience with gRPC or other bidirectional binary streaming protocols.
Experience with audio streaming and related technologies (WebRTC, WebSockets, etc.).
Experience with a multilingual monorepo where you pick the best language out of merit more than personal experience.
Experience with multi-cloud infrastructures (AWS, GCP, OCI, etc.).
Comfort with configuration management tooling (Ansible, Chef, Puppet, or similar).
SRE, DevOps, or platform engineering background at a startup.
Experience at an early-stage company.

Responsibilities

Architecture and implementation of Rime's TTS serving infrastructure, from GPU-backed inference engines to the API surface.
Model optimization from a single-node to disaggregated fleet serving.
Compatibility with different NVIDIA hardwares from Hopper to Blackwell and beyond for on-prem and cloud deployments.
Continuous integration and deployment workflows for the model serving pipeline.
Site reliability: on-call rotation, monitoring, alerting, and observability across the serving stack.
Resource provision, cost management across our GPU fleet.