Voice AI - Solution Architect

Job Description: Voice AI Solutions Architect - ICICI Lombard

About ICICI Lombard:
ICICI Lombard is a leading insurance company known for its digital prowess and innovative use of technology. The company embraces digital transformation and employs cutting-edge technologies such as AI, ML, and big data analytics to provide exceptional insurance experiences. By investing in advanced digital platforms and tools, ICICI Lombard simplifies insurance processes, facilitates customer interactions, and offers user-friendly mobile apps and web portals for policy purchase, claims filing, and self-service options.

ICICI Lombard remains at the forefront of the digital landscape by actively exploring emerging technologies to unlock new possibilities in the insurance industry. By harnessing data and analytics, the company develops personalized insurance solutions and effectively manages risks. Internally, ICICI Lombard optimizes operations, automates processes, and encourages collaboration through a tech-driven approach.

With its unwavering commitment to digital innovation and technological advancements, ICICI Lombard is poised to transform the insurance industry and redefine insurance service delivery, establishing itself as a digital insurance leader.

About the Role

We are seeking a Voice AI Solutions Architect to design, build, and scale our next-generation telephony-integrated voice capabilities. Unlike a pure product role, this position requires a strong technical architect who can conceive and deploy end-to-end voice AI frameworks that sit at the intersection of traditional telephony (SIP/VoIP) and modern Generative AI.

You will act as the technical bridge, working closely with AI teams to operationalize Large Language Models (LLMs) within real-time voice channels. You will be responsible for the architectural integrity, latency optimization, and reliability of voice agents that handle complex customer interactions.

Band

L5-L7

Experience

4-6 years

Vertical

Data Science & Analytics

Location

Mumbai

Reporting to

Program Manager/Sr. Program Manager – Data Science

Responsibilities:

• Architect Telephony-Embedded Voice Systems: Design and deploy end-to-end voice AI frameworks that integrate seamlessly with enterprise telephony systems (e.g., Genesys, Avaya, Exotel, etc.) and cloud communication platforms.
• Low-Latency System Design: Engineer real-time pipelines for Speech-to-Text (STT), LLM inference, and Text-to-Speech (TTS) to achieve sub-second latency in voice conversations.
• AI Team Collaboration: Work closely with the Data Science and Cloud Engineering teams to select, fine-tune, and implement the right models (LLMs/SLMs) for voice contexts, ensuring models are optimized for conversational speed and accuracy.
• Integration Strategy: Manage complex integrations between voice gateways (SIP trunks, WebSockets), orchestration layers, and backend policy systems to ensure fluid data exchange during calls.
• Reliability & Scalability: Define best practices for monitoring call quality, handling interruptions (barge-in), and ensuring high availability of voice agents during peak loads.
• Technical Leadership: Lead the technical vision for the voice platform, maintaining documentation and testing frameworks for voice specific nuances (accents, background noise, turn-taking)

Educational Qualifications –

• B.E/B.Tech, M.E/M.Tech from premium institutes.
• MBA or relevant certification in System Architecture is a plus.

Competencies Required:

Experience -

• Experience: 4-6 years of total experience with a significant focus on Voice AI, Telephony, or Conversational Architecture.
• Telephony & Voice AI: Proven track record in conceiving, developing, and deploying end-to-end telephony-embedded voice AI frameworks. Deep understanding of SIP, RTP, WebSockets, and integration with contact center platforms (CCaaS).
• AI/ML Interaction: Demonstrated experience working closely with AI/ML teams to deploy models in production. Understanding of LLM constraints, context caching, and token streaming for voice.
• Cloud Native: Strong experience with cloud deployments (Azure/AWS/GCP) and containerized voice services (Kubernetes/Docker).
• Coding: Strong fundamentals in Python, Node.js, or Go, specifically for real-time applications.

Skills Required:

Skills

Medium - Advanced

Expert

Voice Architecture

Integration of STT/TTS with basic IVR flows.

Full-duplex architecture; sub-second latency optimization; handling complex "barge-in" & interruption logic.

Telephony Integration

SIP Trunking basics; CTI integration.

Deep integration with Genesys/Avaya/Exotel; WebSocket streaming for raw audio; VAD (Voice Activity Detection) tuning.

Gen AI for Voice

Prompt engineering for chat.

Optimizing LLMs for voice-first interactions (brevity, tone); RAG for real-time call guidance.

Protocol Knowledge

HTTP/REST APIs.

RPC, MRCP, SIP, WebSocket, RTP/SRTP.

Behaviour Required –

OneIL1Team -

• Growth Mindset: Constantly exploring new voice technologies (e.g., end-to-end audio models).
• Team Player: Thrives in a collaborative environment between infrastructure, telephony, and data science teams.
• Adaptability: Agile enough to pivot strategies as Generative AI voice capabilities evolve rapidly.