At [24]7.ai, I develop a distributed voice adapter that connects LLM-based AI agents with Azure Communication Services. I implemented DTMF flows, no-input timers, and WebSocket connection optimizations for real-time voice interactions.
I also re-engineered the Agentic Platform Adapter from a stateful architecture to a stateless one with Redis-backed session management. This enabled scalable Kubernetes deployments with zero downtime.
Previously, I migrated core services to Microsoft Azure and modernized the Speech Platform to handle 100K+ daily calls while maintaining backward compatibility. I also built a Kafka-based streaming system processing 100K events per second and hold the AZ-900 Azure Certification.
