Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

Skit AI

Lead DevOps Engineer

Posted Updated

Description

Responsibilities:

  • Own production infrastructure across AWS, GCP, and Azure, including private interconnects, transit and hub-and-spoke topologies, and cross-cloud latency management with explicit p95 budgets.
  • Design and operate self-hosted LiveKit and SIP infrastructure at scale, including session-affine, event-driven autoscaling for stateful media servers that can handle rapid traffic growth.
  • Build and manage model-serving infrastructure for GPU fleets supporting self-hosted ASR and open-weight LLMs, including inference optimization, prefix caching, sticky-session routing, sub-500ms TTFT targets, and managed API integrations across Vertex AI, Bedrock, and Azure.
  • Lead reliability and observability initiatives using OTel-native tracing, latency attribution across telephony, ASR, LLM, and TTS, automated incident response, and self-healing systems.
  • Act as incident commander for infrastructure incidents and create practical runbooks for production operations.
  • Drive cost engineering with cost-per-minute as an SLO through caching, rightsizing, autoscaling, and workload re-architecture.
  • Implement Zero Trust security across clouds, including private endpoints, IAM/RBAC, secrets rotation, and WAF/DDoS protections.
  • Operate and maintain controls for SOC 2 and ISO/IEC 27001, and support audits involving banking, telecom, and data residency requirements.
  • Provide technical leadership through Terraform-first IaC standards, production-readiness reviews, and mentoring SREs.
  • Tackle current challenges, including scaling stateful self-hosted media servers beyond current concurrency limits, migrating LLM inference from managed APIs to self-hosted GPU-based open-weight models without breaking TTFT budgets, optimizing cross-cloud topology for ASR, LLM, and telephony, and building multi-region disaster recovery without doubling spend.


Requirements:

  • 6+ years of hands-on cloud infrastructure experience.
  • 3+ years operating multiple clouds simultaneously in production, with deep expertise in at least two of AWS, GCP, and Azure.
  • Production experience with real-time audio or video systems such as WebRTC, SIP/PSTN, or streaming media, including debugging jitter and related issues.
  • Strong networking expertise in VPC/VNet design, load balancing, DNS, NAT, private connectivity, transit gateways, and cross-cloud mesh architectures.
  • Experience with Kubernetes at scale across EKS, GKE, or AKS, Helm, and scaling stateful workloads.
  • Familiarity with service mesh technologies such as Istio or Linkerd.
  • Strong Infrastructure as Code experience with Terraform across multi-account or multi-project environments.
  • Production AI/ML serving experience, including GPU allocation and scheduling, inference servers or serverless GPU platforms, and streaming protocols such as WebRTC, WebSocket, and gRPC.
  • Experience operating production STT, TTS, and LLM APIs, including streaming integrations, quota management, and multi-vendor failover.
  • Strong security fundamentals, including IAM/RBAC, secrets management, encryption, and key rotation.
  • Experience with CI/CD using GitHub Actions or GitLab CI, with integrated security scanning.


Nice to Have:

  • Experience with LiveKit, pipecat, Twilio, or similar real-time platforms.
  • Experience with SIP trunking and PSTN integration.
  • Hands-on use of KEDA or other event-driven autoscaling in production.
  • MLOps experience, including model versioning and canary or A/B rollouts.
  • FinOps experience, including reserved or spot strategy and unit economics reporting.
  • Certifications such as AWS Solutions Architect Professional, GCP Professional Cloud Architect, or Azure Solutions Architect Expert.
  • Exposure to ISO/IEC 42001:2023.

Skills

What Lead DevOps jobs ask for — and how much of it you have →
Apply

See also

DevOps jobs by country — openings, pay and top skills →