GenAI Lead

GenAI Lead Engineer– Job Description Role Overview: We’re looking for a GenAI lead to design, build, and deploy intelligent LLM-powered systems—from single-agent chatbots, copilots to complex multi-agent applications—at scale.We are particularly interested in candidates who have hands-on experience in taking GenAI applications from concept to production, especially within high-volume B2C environments. This role prioritizes individuals who understand the nuances of deploying, maintaining, and optimizing GenAI solutions for real-world users, beyond the scope of Proof-of-Concept (PoC) development.You will work across the full stack, integrating LLMs, microservices, vector databases, backend APIs, and modern cloud infrastructure.Key Responsibilities: 1. GenAI Application Development & Deployment Develop scalable, asynchronous microservices usingPython (FastAPI)for chatbots, copilots, and agentic workflows. Design event-driven architectures to supporthigh concurrency , rate limiting, andreal-time responsiveness . Implement secure, versioned REST/gRPC APIs UsePydantic , dependency injection, and modular coding practices for maintainability. Proficient in working with databases usingORMs like SQLAlchemy Ensure observability using logging, metrics, tracing, and health checks. Create responsiveReact.jsfrontends integrated via REST APIs or WebSockets. Deploy applications onCloud Run ,GKE , using Docker, Artifact registry, CI/CD pipelines2. LLM-Powered Conversational Interfaces Design and build LLM-poweredchatbots, voicebots, copilots and other applicationsusing LangChain or custom orchestration frameworks. Integrateenterprise-grade LLM APIs(Gemini, OpenAI, Claude) for multi-turn, intelligent interactions. Implementuser session managementandcontext/state trackingfor personalized and continuous conversations. BuildRAG pipelineswith vector databases, knowledge graphs to ground responses with external knowledge and documents. Applyadvanced prompt engineering(ReAct, Chain-of-Thought with tool calling) for precise and goal-oriented outputs. Ensure performance inlow-latency, streaming environmentsusing WebSockets, gRPC, and SIP media gateways. Performfine-tuning of open-source LLMs(LLaMA variants) using techniques like SFT, LoRA, for cost-effective domain adaptation. Optimizehigh-speed inference pipelinesleveragingmulti-GPU clusters (up to 8x H100s)to reduce latency and improve throughput. 3. Multi-Agent Systems & Orchestration Createmulti-agentsystems & Implementorchestrationpatterns like supervisor-agent, hierarchical, and networked agents using frameworks like ADK, Pydantic AI and LangGraph. UseLangGraphfor stateful workflows with memory, conditional branching, retries, and async execution. Enable persistent context and long-term memory Monitor behavior, drift, and performance using observability tools. Skilled in developing agents with ADK and A2A protocols & experienced in configuring custom and remote MCP servers.Preferred Tech Stack: Languages/Frameworks:Python, FastAPI, HTML, CSS, React.js, LangChain, LangGraph, Pydatic AI, ADK (Agent Development Kit) LLMs & Agents:OpenAI (GPT-4), Claude, Gemini, Mistral, LLaMA 3.2/4 Databases:BigQuery, Redis, FAISS, Pinecone, SQLAlchemy, Chroma, GCP Vector search Protocols/APIs:REST, gRPC, WebSockets, OAuth2, OpenAPI, MCP, A2AAdditional Good to have Tech Stack: DevOps:Docker, GitHub Actions, Jenkins, GKE, Cloud Run Infra & Tools:GCP, Azure, Pub/Sub,Artifact Registry, NGINX, Langfuse, Postman, PytestYears of Experience: Overall Experience: 10 + years of experience (with 2+ years in GenAI) Location: - Pune