Location - Bangalore, Hyderabad, Gurgaon
Years of experience - 4Years+
Mode of work - Onsite
Notice Period - Immediate to 30days

Role Overview
As an AgenticAI Engineer - L2 at CoffeeBeans, you will play a pivotal role in shaping the future of AI workflows. You will leverage your expertise in LangGraph and LangChain to build sophisticated agentic workflows that not only enhance signal correlation and contextual reasoning but also drive actionable insights. This is an unparalleled prospect to be at the forefront of AI innovation, transforming how organizations harness the power of data and AI.
Key Responsibilities
• Design and implement production-grade agentic workflows using LangGraph and LangChain for advanced signal correlation and contextual reasoning.
• Develop a robust memory and context layer to manage workflow states, historical context, and knowledge retention.
• Create an AI engineering harness to streamline prompt management, model configuration, and workflow testing with built-in guardrails.
• Establish evaluation frameworks to assess response quality, signal relevance,



and workflow accuracy, ensuring optimal tool usage.
• Implement comprehensive observability for agent traces, including state transitions, latency, and error tracking to uphold production quality.
• Develop secure APIs and reusable AI services utilizing Python and FastAPI for seamless integration.
• Integrate agents with event streams, enterprise data sources, APIs, and vector stores to enhance application interoperability.

Required Qualifications
• 2+ years of experience building production-grade agentic or LLM applications using LangGraph/LangChain.
• 2+ years in memory and context engineering, focusing on state management and contextual retrieval.
• 2+ years of experience in agent evaluation, including offline evaluations and regression testing.
• 3+ years designing schemas and queries in MongoDB or similar document databases.
• 4+ years of experience building secure REST APIs and

📌 Agentic AI Engineer (India)
🏢 CoffeeBeans
📍 India