Tavus
Software Engineer, Infrastructure
Remote +1 · Hybrid · Full-time · Posted Apr 20, 2026
- Location
- Remote
San Francisco - Experience
- No specific years stated; requires a senior track record of ownership.
About the job
Own and improve the GPU infrastructure and supporting systems that run Tavus’s real-time conversational product, focusing on uptime, latency, throughput, and cost.
What you'll do
- Manage and optimize GPU inference deployments across providers and regions, including cluster setup, routing, scheduling, and model loading.
- Improve system uptime and security, including SOC2 work, and fix issues across infrastructure, backend services, and product code.
- Improve deployment pipelines and GPU performance, including cold starts and model load times.
What they ask for
- Hands-on experience deploying and optimizing GPU inference workloads and building reliable systems with GPU cloud providers.
- Deep Kubernetes and EKS experience, including routing and scheduling.
- Deep AWS experience, including creating services and repeatable processes.
- Senior-level ownership: setting technical direction, making decisions, completing ambiguous work, and communicating clearly.
Nice to have
- Experience with GCP, video streaming infrastructure, training infrastructure, or LLM serving.
- Experience with SOC2 or security compliance.
Tools
GPU, Kubernetes, EKS, AWS, GCP, CUDA, SOC2, LLM
AI focus
Operate and optimize GPU inference infrastructure for real-time conversational AI, including model serving and video rendering.
Model-written summary · the original posting is authoritative