T

Telehealth Platform

Telehealth

Case Study

Video Conferencing Agents & AI Integration Using LiveKit

Scalable telehealth platform integrating human agents, AI automation, and real-time video — handling thousands of daily consultations on Kubernetes.

LiveKitKubernetesWebRTCAI AgentsPrometheusGrafana
5,000+
Concurrent sessions supported
<100ms
Avg end-to-end latency
99.9%
Platform uptime
2 regions
Global edge deployment

The Challenge

The client needed a reliable, HIPAA-adjacent platform capable of routing consultations to either human agents or AI-driven workflows in real time, with no perceptible handoff latency and consistent video quality across low-bandwidth connections.

Our Solution

KubeAce architected a LiveKit-based SFU deployment on Kubernetes, integrating LiveKit Agents for AI pipeline orchestration. We built room-level routing logic, Egress pipelines for session recording, and a full WebRTC observability stack to monitor media quality in production.

Overview

The client operates a telehealth platform serving patients across multiple geographies. Consultations are handled by either a human clinical agent or an AI-assisted workflow depending on triage — requiring a video infrastructure that could support both paths seamlessly.

Challenge

The platform needed real-time video capabilities that could:

  • Scale to thousands of simultaneous sessions during peak hours
  • Support AI agent integration without disrupting patient-facing audio/video quality
  • Provide session recording for compliance and audit purposes
  • Operate reliably across variable network conditions

Solution

KubeAce deployed a production LiveKit SFU cluster on Kubernetes (GKE), with:

  • LiveKit Agents for server-side AI pipeline integration — handling transcription, triage routing, and AI voice responses without client SDK changes
  • Multi-region deployment with latency-based routing for sub-100ms session establishment
  • Egress pipelines for composite recording to GCS, triggered per session type
  • WebRTC observability — packet loss, jitter, and MOS score dashboards in Grafana with alerting thresholds before quality degradation becomes user-visible

Results

The platform handled peak load from day one, with no unplanned downtime in the first 90 days post-launch. AI-assisted consultations run on the same infrastructure as human-agent sessions with no architectural distinction from the client SDK perspective.

More Case Studies

View All →
Free 30-Minute Strategy Session — No Commitment

Ready to Transform
Your Infrastructure?

Whether you're migrating to Kubernetes, scaling a LiveKit deployment, or building a DevOps platform from scratch — our engineers have done it before.

Schedule a Free Call View Case Studies
Response within 4 hours
NDA available on request
Serve clients globally from Bangalore