Required Skills
About the Job
JLL Technologies is seeking an experienced Senior Software Engineer specializing in Agentic AI to join our Innovation engineering team in Bengaluru/Bangalore. In this role, you will design, build, and deploy production-grade multi-agent AI systems that enhance our Azara AI-driven data intelligence platform for commercial real estate. You'll architect autonomous agent workflows, build RAG pipelines, orchestrate LLM interactions, and deliver scalable AI solutions that create significant business value.
Key Responsibilities: * Design and develop production-ready multi-agent systems using LangGraph, with experience in frameworks like LangChain or CrewAI. * Build and optimize RAG pipelines, including document processing, embedding, and vector database integration. * Integrate and manage LLM services (OpenAI, Azure OpenAI, Anthropic) with a focus on prompt engineering and cost optimization. * Develop scalable Python backend services (FastAPI) and event-driven microservices for AI agent workflows. * Implement distributed task processing (Celery) and autoscaling for AI workloads. * Architect and implement agent orchestration patterns, including planning, tool use, persistent state, and memory. * Lead proof-of-concept development and mentor engineers on AI engineering best practices.
Required Qualifications: * Strong Python proficiency with production AI application experience. * Demonstrated experience with LangGraph or similar agentic AI frameworks. * Hands-on experience with LLM API integration and prompt engineering. * Experience designing and implementing RAG systems. * Solid understanding of multi-agent system design and orchestration. * Experience with FastAPI and Celery. * Proficiency with Git, CI/CD, and cloud platforms (preferably Azure). * Experience with AI-powered development tools is a plus.
Preferred Qualifications: * Experience with vector databases (e.g., PgVector). * Experience with Databricks Genie or similar natural language to data query platforms. * Experience with containerization (Docker, Kubernetes) and event-driven autoscaling (KEDA).
