Voice AI Agent Product Manager

VOICE AI AGENT — PRODUCT MANAGER

Location: Noida (on-site)

Type: Full-time | Client-facing

Experience: 3.5+ years minimum as an AI Product Manager

Compensation: Up to ₹25 LPA fixed

ABOUT THE ROLE

We build AI employees — always-on conversational agents that run core business functions such as sales, customer support, hiring, and growth — and we apply the same technology to healthcare, across clinical workflows, medical education, and patient engagement.

We are looking for a Product Manager to own voice AI agents end to end: from discovery and requirements through launch, and then through the long tail of tuning that decides whether an agent actually works in production.

This is a hands-on role. You will spend as much time inside prompts, call recordings, and evaluation data as you will in roadmap documents.

It is also a client-facing role. You will sit across the table from enterprise and healthcare customers — running discovery, demoing agents, negotiating scope, and translating what you hear into a build plan engineering can execute.

WHAT YOU'LL OWN

Product direction

• Define vision, roadmap, and success metrics for voice agents across enterprise and healthcare use cases

• Write clear PRDs: conversation scope, intents, escalation paths, guardrails, integrations, and acceptance criteria

• Own and groom the backlog; make prioritisation calls and defend them with data

• Coordinate releases across engineering, data science, design, and operations

Agent quality

• Design and iterate agent behaviour: system prompts, dialogue flows, tool/function calling, knowledge grounding, fallback and human-handoff logic

• Build evaluation sets and run structured testing against them; drive measurable gains in accuracy, containment, and responsiveness

• Listen to real calls. Tag the failures, find the pattern, turn it into a fix.

Client-facing delivery

• Run discovery workshops and requirement sessions with client teams

• Demo agents, handle objections, and set honest expectations on scope, timelines, and what the technology can and cannot do

• Own the feedback loop from pilot through production rollout

Data and metrics

• Track and improve the metrics that matter in voice: ASR accuracy / WER, intent resolution, containment and deflection rate, end-to-end latency and barge-in responsiveness, escalation rate, hallucination and off-script incidents, CSAT

• Use call analytics, user research, and experiment results to decide what ships next

Compliance

• Ensure voice solutions meet applicable regulatory and data-protection requirements, particularly in healthcare — consent capture, PHI handling, recording and retention, auditability

• Work with engineering and legal to build compliance into the product rather than bolt it on

MUST-HAVES

• 3.5+ years minimum as an AI Product Manager. Total PM experience may be higher, but at least 3.5 years must be hands-on ownership of AI/ML or conversational AI products.

• Direct experience with voice or conversational AI — ASR/STT, TTS, NLU, LLM-based agents, or IVR-to-AI migrations

• Strong, demonstrable prompt engineering ability — you will be tested on this (see hiring process below)

• Working understanding of the voice stack: telephony/SIP or WebRTC, streaming ASR, turn-taking and interruption handling, latency budgets — and where each one breaks

• Track record of shipping with cross-functional teams and managing stakeholder expectations

• Strong analytical and problem-solving skills; comfortable defining metrics, reading dashboards, and running experiments

• Excellent communication — you will write specs engineers can build from and present the same thing to non-technical client stakeholders

• Experience with agile methodologies, product discovery, and iterative development

• Comfortable in a fast-paced, evolving environment, with genuine ownership of outcomes

• Bachelor's degree in engineering, computer science, business, or a related field

GOOD TO HAVE

• Healthcare or health-tech exposure — clinical workflows, patient engagement, regulatory context (HIPAA, DPDP)

• Hands-on with voice agent platforms and orchestration frameworks (LiveKit, Pipecat, Vapi, Retell, Twilio, or equivalent in-house stacks)

• Experience building LLM evaluation harnesses or LLM-as-judge pipelines

• Multilingual or Indic-language voice deployments

• Advanced degree (MBA / MS)

HIRING PROCESS

1. Recruiter screen — background, experience mapping, expectations

2. Prompt engineering test — a practical exercise: design and refine the prompt and behaviour spec for a voice agent. Define scope and persona, handle edge cases and interruptions, prevent hallucination, and specify escalation. Scored on structure, edge-case coverage, and how you reason about failure modes — not on wordsmithing.

3. Product deep-dive — walk us through a voice or AI product you owned: what shipped, what the metrics did, what you got wrong

4. Cross-functional round with engineering and data science

5. Client-scenario round — a live discovery and demo simulation

6. Leadership round and offer

DETAILS

Location: Noida — on-site

Employment: Full-time

Compensation: Up to ₹25 LPA fixed

Experience bar: 3.5 years minimum as an AI PM

Role type: Client-facing

THIS ROLE IS PROBABLY NOT FOR YOU IF

• Your AI product experience is limited to writing specs for models someone else evaluated

• You have run roadmaps but never opened a prompt, a transcript, or an eval set

• You would rather not be in front of customers