Lead Agentic AI Engineer – Vice President

Citi is seeking a Lead Agentic AI Engineer at the Vice President level to architect and deliver production-grade agentic AI systems that sit at the intersection of large language models, retrieval-augmented generation, and modern full-stack engineering. This role operates within Citi's technology organization and carries direct ownership over the design, development, and deployment of intelligent, multi-agent platforms that drive measurable outcomes across the business. You will build and scale multi-tenant agentic ecosystems, develop custom fine-tuning applications, and ensure the platform adheres to our core principles: Trust, Adoption, Cost, Operations, and Scalability. You will lead by example, setting the standard for code quality and architectural purity while driving high-impact business use cases. This role requires deep expertise in Python, LLMs, and Agentic workflows, combined with strong full-stack capabilities (Java, React, and microservices) to seamlessly integrate AI capabilities into enterprise-grade platforms.

1. Core Platform & Application Development Full-Stack Engineering: Lead the development of custom AI platform components. Build and maintain full-stack applications utilizing Python, Java (Spring Boot), and modern frontend frameworks like React (e.g., developing internal LLM/SLM fine-tuning planes, microservices, and experiment tracking dashboards). Agentic Frameworks: Code and optimize multi-tenant intelligent agents utilizing modular orchestration patterns such as ReAct and ReWOO, and frameworks like Google's Agent Development Kit (ADK). Strict Architectural Implementation: Develop and enforce clean, decoupled integration layers. Build Model Context Protocol (MCP) servers ensuring a strict communication flow: Agents interact solely with MCP servers, and MCP servers interact solely with APIs to retrieve data. Direct database access from agents or MCP servers is strictly prohibited. 2. Advanced Data Retrieval & Logic Engineering Next-Generation RAG: Write the data ingestion and retrieval code for advanced Retrieval-Augmented Generation (RAG) architectures, including Knowledge Graph RAG (GraphRAG), LightRAG, and hierarchical summary trees (RAPTOR). Vector & Graph Integrations: Develop seamless integrations with graph and vector databases (such as Neo4j and pgvector) to power complex, thematic data retrieval. Prompt & Intent Engineering: Design robust LLM instructions and classification logic to prevent collisions in complex workflows, ensuring mutually exclusive intents are handled with high precision. 3. Use Case Development & Forward Deployment SME Collaboration: Act as a Forward Deployed Engineer (FDE), working directly with Subject Matter Experts (SMEs) on business and domain understanding to accurately translate complex enterprise workflows into automated, agent-driven code. Seamless Integration: Partner with UI and workflow integration developers to ensure the backend agentic logic connects flawlessly with user-facing layers and existing enterprise APIs.

Required Experience & Education 10+ years of professional software engineering experience. 5+ years of experience in AI/ML, with at least 2+ years specifically in Generative AI (LLMs, RAG, Agentic workflows). 3+ years of leadership experience managing technical teams and delivering complex software or AI solutions. Bachelor’s or Master’s degree in Computer Science, Engineering, or a related technical field. Technical Skills AI/ML & Python: Deep hands-on expertise in Python, LLM orchestration, prompt engineering, and agentic frameworks (ADK, LangChain, LangGraph, etc.). Full-Stack & Microservices: Strong proficiency in backend development using Java (Spring Boot) or Python (FastAPI), and frontend development using React. Data & Databases: Experience integrating applications with relational database like Oracle and Graph databases (Neo4j) and Vector databases (pgvector, Pinecone, or Milvus).