Compensation: ₹20L – ₹25L • No equity

Voice AI Engineer

About timepay.ai

timepay.ai builds AI voice agents and WhatsApp and SMS automation for businesses across industries. Our platform runs lead generation, sales, customer support, retail and collections conversations at scale.

The role

You will build the AI agents our customers' customers actually talk to: voice agents on phone and web calls, WhatsApp agents and chat agents. You will design the conversation flows, wire together STT, LLM and TTS in real time, and make every agent sound natural, respond fast and get the job done.

What you'll work with

• Voice frameworks: Pipecat, LiveKit Agents
• Real-time transport: WebSockets, WebRTC, SIP / telephony
• AI models: STT, LLMs, TTS (in-house and commercial)
• Messaging: WhatsApp Business API, web chat
• Backend: Python, Node.js, PostgreSQL, Redis

What you'll do

Build voice AI agents

• Build and ship real-time voice agents with Pipecat and LiveKit Agents, connecting STT, LLM and TTS into low-latency streaming pipelines.
• Handle the hard parts of live conversation: voice activity detection, turn-taking, barge-in, silence, background noise and call transfers.
• Connect agents to phone networks and browsers through SIP trunks, telephony providers, WebRTC and WebSockets.
• Bring end-to-end response latency down by streaming every stage, choosing the right models and tuning infrastructure.

Build WhatsApp and chat agents

• Build WhatsApp agents on the WhatsApp Business API: templates, session messages, media, buttons and lists.
• Build web and in-app chat agents that share logic with voice agents across channels.
• Design hand-offs between channels and to human agents.

Design conversation flows

• Turn customer use cases (lead qualification, sales, support, reminders, collections) into voice and chat flows.
• Write and tune prompts, system instructions and guardrails; build tool and function calls that let agents look up data, book slots or update CRMs.
• Support multiple Indian languages and code-mixed speech, with natural, on-brand voices.
• Build reusable flow templates and a flow builder so new agents launch fast.

Test, measure and improve

• Build testing and evaluation for agents: simulated calls, transcript review, task success, latency and drop-off.
• Use call recordings, transcripts and analytics to find failures and fix them.
• Store and query conversation state, context and outcomes in PostgreSQL and Redis.
• Work with forward deployed, backend and ML engineers to take agents live for customers and improve them over time.

What success looks like

• 30 days: Learned our voice and messaging stack; built and demoed a working voice agent and a WhatsApp agent end to end.
• 60 days: Taken a customer agent live; cut latency or improved task success on an existing agent, with numbers to show it.
• 90 days: Shipped reusable flow templates or tooling that speed up new agent launches; set up an evaluation loop for agent quality.

What we're looking for

• 2+ years of software engineering, with hands-on experience building voice AI agents, WhatsApp agents or chat agents in production.
• Hands-on experience with Pipecat and/or LiveKit Agents.
• Strong understanding of real-time systems: WebSockets, WebRTC, audio streaming, codecs and latency.
• Experience integrating STT, LLM and TTS providers (such as Deepgram, Whisper, OpenAI, Gemini, ElevenLabs, Cartesia or similar).
• Experience designing voice and chat flows, prompt engineering and LLM tool/function calling.
• Strong Python; comfortable with async programming.
• Working knowledge of databases (PostgreSQL, Redis) for conversation state and analytics.
• Experience with the WhatsApp Business API or other messaging platforms.
• Ownership and a strong ear for conversation quality: you notice when an agent sounds robotic or slow, and you fix it.

Good to have

• Telephony: SIP, Twilio, Exotel, Plivo or similar; Asterisk or FreeSWITCH.
• Multilingual and Indian-language voice agents, including code-mixed speech.
• Node.js and Docker; deploying real-time services on GCP or AWS.
• RAG, knowledge bases and agent frameworks (LangChain, LlamaIndex or similar).
• Building agent evaluation, testing or observability tools.
• Experience with conversational AI for sales, support, retail or BFSI use cases.

Why join timepay.ai

• Build AI agents that hold millions of real conversations across industries and languages.
• Work at the cutting edge of real-time voice AI, with the freedom to try new models and frameworks.
• See your work live fast: agents you build go in front of customers in weeks, not quarters.
• Competitive salary and flexible work.