← All roles

Software Engineer, Infrastructure

Tavus· Engineering, Product, & Design
Apply on company site
Hybrid (Remote), Hybrid (San Francisco)Full-timePosted 6mo ago

About the role

Senior Software Engineer (Infrastructure) to own the systems behind CVI, Tavus' real-time conversational product. Focus on uptime, latency, and cost across GPU inference deployments, multi-provider/multi-region routing, and scalable infrastructure.

Responsibilities

  • CVI's inference deployments: GPU infrastructure serving live conversations across providers and regions, including tuning GPU generations and reducing cold-start/model load times.
  • Expanding our GPU footprint: bring on new providers and regions, stand up clusters on EKS, build routing, scheduling, and throughput for fast weight loading.
  • Uptime: contribute to higher uptime, alongside security and SOC2 work to keep infrastructure trustworthy.
  • Fix what you find: fix or flag issues, rework deploy pipelines to ship faster and more reliably.

Qualifications

  • Hands-on GPU inference experience: deployed and optimized inference workloads on GPUs and built reliable systems on GPU cloud providers.
  • Kubernetes and EKS depth, including routing and scheduling: design how work is placed across a fleet and write the services to enable it.
  • Deep AWS experience: capable of spinning up new services and turning them into repeatable processes for others.
  • A senior track record of ownership: set technical direction, make decisions, and carry ambiguous work to completion; able to explain complex ideas clearly.

Skills

Required:GPU inference deploymentKubernetes / EKSAWS
Last updated Jul 15, 2026. Interested? Apply on the company site.
Tavus logo

About Tavus

Meet the PALs

Tavus introduces PALs: AI humans that remember, empathize, and grow with you. They move fluidly between chat, voice, and video, making AI finally feel human.

AIVideoConversational AIMultimodalAPIsResearch
Company profileWebsite

More open roles at Tavus