•Position – Generative AI Architect
•Experience – 10+ Years
•Location – Pune/ Bangalore/ Kolkata/ Indore
•Skills - Gen AI Development Machine Learning, Retrieval Augmented Generation (RAG), Langchain, Vector Databases, Langgraph, Arize Phoenix, RAGAs, LLM-as-a-Judge, MCP
Notice Period – Immediate to 15 Days
Job Description
•Overview We are seeking an exceptionally skilled and experienced Generative AI Architect to lead the design, development, and deployment of cutting-edge Generative AI solutions. This role requires a blend of deep machine learning and natural language processing expertise with hands-on architectural experience in building scalable, production-grade LLM applications. The ideal candidate will be a technical leader who can drive innovation, define best practices, and mentor engineering teams in the evolving GenAI space.
Key Responsibilities
•Design and lead the implementation of end-to-end, high-performance Generative AI architectures, focusing on scalability, cost-efficiency, and resilience.
•Architect and optimize Retrieval-Augmented Generation (RAG)



pipelines for enterprise-specific knowledge bases, including vector store integration, indexing, and retrieval optimization.
•Design and implement complex, multi-step Agentic AI systems and sophisticated conversational flows using frameworks like LangChain and LangGraph.
•Establish robust MLOps and GenAI-Ops practices, integrating observability and evaluation tools such as LangSmith, Phoenix, or Langfuse to monitor model performance, latency, cost, and drift in production.
•Define and implement LLM evaluation methodologies, utilizing tools like Ragas to quantitatively assess the quality (e.g., faithfulness, answer relevance, context adherence) of RAG and Agentic AI applications.
•Continuously evaluate recent LLMs, foundation models, open-source frameworks, and emerging GenAI technologies to recommend and integrate the best-fit solutions.
•Work closely with Data Scientists, ML Engineers, and business stake

📌 GEN AI Architect (India)
🏢 Calsoft
📍 India