Senior Platform / DevOps Engineer (Real-time Media, WebRTC, Edge + Cloud)
Job title: Platform / DevOps Engineer (WebRTC, Edge + Cloud)
Location: Bengaluru (Hybrid/Office)
Employment type: Full-time
Experience: 5–12+ years (flexible for strong fit)

About the role

We’re building and operating a LiveKit-like real-time communications platform (WebRTC) that must scale to millions of calls with edge PoPs for ultra-low latency and multi-region cloud reliability. This is a hands-on, high-ownership role focused on production systems, performance, and resilience.
We’re especially interested in engineers who’ve seen scale in real-time/streaming infra.

What you’ll do

Own reliability and performance of signaling, SFU/media nodes, TURN, routing, failover, and capacity planning
Build and run multi-region Kubernetes platforms with secure networking and zero-downtime deployments
Design edge + cloud architecture: PoPs, global routing, failover, autoscaling, DR
Implement SLOs/SLIs, incident response, postmortems, and operational excellence
Create strong observability: metrics, logs, tracing, and real-time QoE/latency metrics

• Ship Infrastructure-as-Code and automation: Terraform, Helm, GitOps, CI/CD

Required skills

Strong production experience with Kubernetes at scale (multi-cluster/multi-region)
Strong Linux + networking fundamentals (UDP/TCP, NAT, conntrack, DNS, load balancing)

• Experience with IaC + delivery: Terraform, Helm, GitOps (ArgoCD/Flux), CI/CD
• Proven on-call ownership for high-availability systems

Nice to have

• WebRTC/RTC operations: ICE, STUN/TURN, SFU scaling, packet loss/jitter tuning
• Edge/PoP and traffic management experience (global routing, Anycast/DNS strategies)
• Cost optimization for bandwidth-heavy workloads
• Experience operating realtime/streaming systems at very high concurrency

What success looks like

You can keep a real-time system stable through traffic spikes, packet loss, ISP variability, zone/region failures

• You think in terms of latency budgets, concurrency, bandwidth, packets/sec, not just pods and nodes
• You build platforms that are observable, automatable, and easy to operate