Job
Description
As a Senior AI Engineer specialized in Voice AI and Autonomous Agents at Spyne, you will be responsible for owning and building Spynes in-house voice bot stack. This pivotal role involves leveraging your expertise in LLMs, ASR/TTS, and voice UX to develop immersive, human-like conversations between auto dealerships and their customers. Located in Gurugram, you will work from the office five days a week to drive the development of cutting-edge AI solutions in the automotive retail sector. Your primary responsibilities will include: - Voice AI Stack Ownership: Developing and managing the complete voice bot pipeline encompassing ASR, NLU, dialog state management, tool calling, and TTS to deliver a seamless conversation experience. - LLM Orchestration & Tooling: Designing systems utilizing MCP to facilitate structured context exchange between real-time ASR, memory, APIs, and the LLM. - RAG Integration: Implementing retrieval-augmented generation to support responses based on dealership knowledge bases, inventory data, recall lookups, and FAQs. - Vector Store & Memory: Creating scalable vector-based search functionalities for dynamic FAQ handling, call recall, and user-specific memory embedding. - Latency Optimization: Engineering low-latency, streaming ASR + TTS pipelines and optimizing turn-taking models for natural conversations. - Model Tuning & Hallucination Control: Customizing tone, reducing hallucinations, and aligning responses with business objectives via fine-tuning, LoRA, or instruction tuning. - Instrumentation & QA Looping: Establishing robust observability, conducting real-time call QA processes, and analyzing interruptions, hallucinations, and fallbacks. - Cross-functional Collaboration: Collaborating closely with product, infra, and leadership teams to scale the voice bot solution to numerous US dealerships effectively. To excel in this role, you should possess: - Architectural Thinking: Ability to comprehend the integration of ASR, LLMs, memory, and tools and design modular, observable, and resilient systems. - LLM Tooling Mastery: Proficiency in implementing tool calling, retrieval pipelines, function calls, or prompt chaining across various workflows. - Fluency in Vector Search & RAG: Expertise in chunking, embedding, indexing, and retrieval processes while avoiding prompt bloat and token overflow. - Latency-First Mindset: Capability to identify and rectify token delays, optimize API round-trip time, and ensure human-like call interactions. - Grounding > Hallucination: Skill in tracing hallucinations to weak prompts, lack of guardrails, or tool access deficiencies and addressing them effectively. - Prototyping Skills: Comfort with building from scratch and rapid iteration using open-source or hosted tools as required. Requirements for this role include: - 5+ years of experience in AI/ML or voice/NLP systems with real-time exposure. - Profound knowledge of LLM orchestration, RAG, vector search, and prompt engineering. - Experience with MCP-style architectures and structured context pipelines between LLMs and APIs/tools. - Familiarity with integrating ASR (Whisper/Deepgram), TTS (ElevenLabs/Coqui), and OpenAI/GPT-style models. - Strong understanding of latency optimization, streaming inference, and real-time audio pipelines. - Proficiency in Python, FastAPI, vector DBs (Pinecone, Weaviate, FAISS), and cloud infrastructures (AWS/GCP). - Solid debugging, logging, and QA capabilities for hallucination, grounding, and UX analysis. Join Spyne for a real-world AI impact, a superior team balancing speed with technical depth, high autonomy, and visibility from day one, accelerated career growth, MacBook along with essential tools, and a flat structure focused on meaningful work without unnecessary bureaucracy.,
