Sasi Varunan

CEO

Bengaluru, Karnataka, India14 yrs 9 mos experience
Most Likely To SwitchHighly Stable

Key Highlights

  • 14+ years in Voice AI and Real-Time Communications.
  • Expert in building resilient, low-latency voice systems.
  • Proficient in VoIP protocols and observability strategies.
Stackforce AI infers this person is a SaaS-focused Voice AI Architect with extensive experience in real-time communications.

Contact

Skills

Core Skills

Voice AiVoipVoice EngineeringWebrtcCloud Gaming

Other Skills

ASRAngularJSAsteriskDockerFreeswitchGPU streamingJavaScriptKaiosKamailioKubernetesNode.jsSCTPSIPSTTTTS

About

Voice AI Architect | 14 Yrs in Real-Time Comms ✦ SIP ✦ ASR ✦ LLM ✦ TTS ✦ Observability ✦ Kubernetes/On-Prem Deployments With 14+ years of deep engineering experience, I specialise in building end-to-end Agentic Voice AI systems that are production-grade, real-time optimised, and observability-rich. From SIP-level call routing and SBCs to weaving together ASR, LLM, and TTS pipelines, I build systems that handle millions of sessions with sub-second latency. I’ve worked extensively with SIP proxys, media servers, audio transcoding, TLS/SRTP, and WebRTC clients, deploying them across Kubernetes, Docker Swarm, and bare-metal on-prem environments. Resilience is non-negotiable. I design voice platforms with built-in failover, fallback strategies, and fine-grained observability at every hop — from initial packet to final spoken word.

Experience

Yellow.ai

2 roles

Principal Engineer of Voice Product

Promoted

May 2023Present · 2 yrs 10 mos · Bengaluru, Karnataka, India

  • At my current role, I lead the architecture, development, and scaling of VoIP-powered Agentic Voice AI systems — tightly optimized for real-world use cases, resilience, and ultra-low latency.
  • 🔹 Key Responsibilities 🔹
  • 🔹 VoIP Telephony Solutions
  • Designed and implemented VoIP media infrastructure, including
  • Media Servers (including WebSocket-based audio streaming)
  • Stateful and Stateless SIP Proxies
  • Registrars, Authentication Modules, and Location Servers
  • Enabled secure, scalable SIP-based communication with SRTP/TLS, NAT traversal, and load-aware routing.
  • 🔹 Agentic Voice AI Bot Architecture
  • 🗣️ ASR (Speech Recognition)
  • Architected ASR systems with advanced customization for:
  • Alpha-numeric parsing, domain-specific vocabulary, multilingual support, human name recognition, and boosted keyword phrasing
  • Handled end-of-thought detection, sentiment awareness, VAD (Voice Activity Detection), and cost-saving strategies (like burst detection & segmentation tuning)
  • 🔊 TTS (Text to Speech)
  • Integrated TTS engines that prioritize:
  • Low-latency and natural-sounding voice quality
  • Smart caching strategies for repeat prompts — maintaining variation while reducing compute load
  • 🧠 LLM (Large Language Models)
  • Built custom workflows to:
  • Filter LLM stream output using meaningful delimiters
  • Maintain contextual memory across workflow switches
  • Enhance TTS voice mapping for better expressiveness and relevance
  • 📈 Observability
  • Built end-to-end observability into the voice stack:
  • Latency tracking and performance monitoring at every hop (ASR → LLM → TTS)
  • Real-time alerts, log tracing, and custom metrics dashboards
  • 🔹 Scaling & Reliability
  • Architected High Availability (HA) and auto-scaling infrastructure for VoIP media systems
  • Voice bot services using Docker and Kubernetes
Voice engineeringVoice AIASRSTTTTSVoIP+2

SDE 3 | Voice Engineering

Nov 2021Apr 2023 · 1 yr 5 mos · Bengaluru, Karnataka, India

golangNode.jsKubernetesdockerWebrtcKamailio+4

Jio

Senior Manager - JioGames

Nov 2017Nov 2021 · 4 yrs · Bangalore

  • Led code development, drove projects, and mentored team members within the Jio Games KaiOS/Firefox app development team.
  • Developed an SIP/SCTP-based WebRTC Multiplayer SDK.
  • Conducted R&D and delivered end-to-end MVP for cloud game streaming solutions.
  • Designed and implemented GPU machine-to-user browser streaming (peer-to-peer), allowing users to launch AAA games from the cloud directly within their personal browser.

Plivo

Sr Webrtc Engineer

Dec 2016Nov 2017 · 11 mos · Bangalore

  • Led the development of a WebRTC and SIP-based plugin for browsers, a Chrome extension, and an Electron app.
  • Played a key role in enabling the creation of a full-featured web phone

Talisman telecom

Sr Software Engineer

Mar 2016Dec 2016 · 9 mos · Bangalore

  • 1. End-End development with WebRTC and SIP based Browser SDK,
  • a pluggable call button for any websites.
  • 2. Development of websocket server using NodeJS,
  • Freeswitch/Asterisk VOIP server for PSTN connectivity.

Chitrchatr communications inc.

Senior Voip Engineer (R&D)

May 2015Feb 2016 · 9 mos · Bangalore

  • Research and development on webRTC Javascript client with
  • capabilities of SMS, Fax, Voice and Video calls by using Freeswitch VOIP
  • server infra.

Full creative

Voip Engineer

Mar 2011Mar 2015 · 4 yrs · Chennai Area, India

  • Synclio: Development on VOIP infra setup, Browser phone UI creation with
  • Twilio/Plivo Browser SDKs.
  • “Easy work from home” a solution by replacing hardware VOIP phones with
  • browser based phone with features of 1-1, multiple video calling using peer
  • based webRTC API stack and PSTN calling with browser using jsSIP/Twilio

Education

Anna University Chennai

Bachelor of Engineering - BE — Electronics and Communications Engineering

Stackforce found 100+ more professionals with Voice Ai & Voip

Explore similar profiles based on matching skills and experience