We are hiring a Voice AI Prompt Engineer to own prompt design and agentic workflow engineering across our Voice AI platform. You will build prompts for real-time telephony systems, IVR and outbound bots, multilingual conversational agents, and post-call analytics — as well as multi-step agentic LLM applications. You understand ASR noise, spoken language constraints, and latency budgets, and you work systematically to close the gap between what a model can do and what a voice product needs it to do.
WHAT YOU WILL DO:
- Voice Agent Prompting: Design ASR-noise-robust prompts for IVR bots, outbound calling agents, intent detection, entity extraction, and dialog state tracking within real-time latency budgets
- Spoken Output Design: Optimize LLM responses for TTS delivery — concise, natural-sounding, and telephony-appropriate (no markdown, prosody-aware)
- Multilingual Handling: Engineer prompt strategies for code-mixed input (Hinglish, Tanglish, and other Indian language mixes) with robust fallback logic
- Agentic Pipelines: Build ReAct/chain-of-thought workflows, tool-calling pipelines, and RAG-backed agents for voice and non-voice enterprise applications
- Evaluation & Testing: Build prompt evaluation frameworks with voice-specific metrics; maintain a versioned prompt library with regression tests
- Context Management: Handle long multi-turn voice conversations — summarization, memory injection, and context prioritization strategies
- LLM Integration: Embed prompt logic into production inference pipelines; design structured output schemas for downstream telephony integrations
- Cross-Team Collaboration: Translate product requirements into prompt specifications; curate instruction datasets for fine-tuning from prompt experiments
MUST HAVES:
- Experience 3–6 yrs with LLMs
- 1+ year in voice AI or conversational AI Hands-on experience with ASR output characteristics and downstream LLM effects.
- Experience with Agentintic frameworks like LangChain, LlamaIndex, LangGraph, or equivalent RAG architectures like chunking, retrieval, prompt integration.
- Strong Python experience.
- Experience on evaluation pipelines, API integration, data processing Prompt evaluation with quantitative metrics and regression testing.
- Strong critical reading of model outputs with precise failure mode articulation.
WHY JOIN GNANI:
India’s leading enterprise Voice AI company. 30M+ voice AI calls/day across 40+ languages. Deployed in BFSI, telecom, government, and enterprise. Backed by Samsung Ventures & Info Edge Ventures. Selected under the IndiaAI Mission to build foundational AI models for India.
You will work on live systems processing 30M+ calls/day across BFSI, telecom, and government verticals. Your prompt decisions will have direct, measurable impact on call containment rates and customer outcomes at scale. You’ll have access to proprietary multilingual speech datasets, in-house ASR and TTS models, and real call transcripts across 40+ languages. The role has a clear path into Senior Voice AI Engineer, Voice AI Platform Lead, or LLM Systems Architecture.
