- Build and own the full voice bot pipeline including ASR, NLU, dialog management, tool calling, and TTS.
- Architect systems using MCP to connect ASR, memory, APIs, and LLMs in real-time.
- Implement RAG to ground responses using data from knowledge bases, inventory, and FAQs.
- Design scalable vector search systems for memory embedding and FAQ handling.
- Engineer low-latency ASR and TTS pipelines, optimizing for natural turn-taking.
- Apply fine-tuning, LoRA, and instruction tuning to reduce hallucinations and align model tone.
- Build observability systems and QA pipelines to monitor calls and analyze model behavior.
- Collaborate with cross-functional teams to scale the voice bot to thousands of users.
- Design modular, observable, and resilient AI systems.
- Implement retrieval pipelines, function calls, and prompt chaining across workflows.




- Expertly chunk, embed, and retrieve documents in RAG systems.
- Debug latency issues and optimize for low round-trip time.
- Trace hallucinations to root causes and fix via guardrails or tool access.
- Build prototypes using open-source or hosted tools with speed and flexibility.
- 5+ years in AI/ML or voice/NLP with real-time experience.
- Deep knowledge of LLM orchestration, vector search, and prompt engineering.
- Experience with ASR (Whisper, Deepgram), TTS (ElevenLabs, Coqui), and OpenAI models.
- Skilled in latency optimization and real-time audio pipelines.
- Hands-on with Python, FastAPI, vector DBs, and cloud platforms.

📌 AI Engineer – Voice AI / Autonomous Agents (Gurugram)
🏢 Spyne
📍 Gurugram