Risk Modeling Solutions - Full-stack GenAI - Assistant Vice President

Citigroup Inc.; Bengaluru; On-site; INR 4,000,000 - 7,000,000; Full time

Job summary

Citigroup Inc. in Bengaluru is seeking a seasoned Full-Stack Gen AI Lead Engineer to architect agentic AI systems, design end-to-end AI workflows, and deliver scalable AI solutions for risk modeling.

You will build backend Python microservices with FastAPI, implement RAG pipelines, containerize with Docker, deploy on Kubernetes, and ensure observability with Langfuse while advancing model governance and AI safety.

Qualifications

• 7+ years of professional experience in software development and data science/ML.
• Expert-level Python development and building scalable backend services (FastAPI preferred).
• Hands-on experience with multiple agentic frameworks (LangChain, LangGraph, CrewAI, AutoGen, Semantic Kernel).
• Experience architecting RAG systems and vector DBs (OpenSearch, Pinecone, Weaviate).
• Experience deploying containerized AI systems on AWS/Azure/GCP.
• Knowledge of MLOps, CI/CD, observability, and evaluation for LLM-based systems.
• Strong understanding of AI risk, safety, governance.
• Background in ML, DL, NLP, Transformer architectures.

Responsibilities

• Architect and lead multi-agent AI workflows with advanced reasoning and autonomous execution.
• Translate business risk problems into robust AI solutions.
• Develop Python microservices and REST APIs exposing AI capabilities.
• Implement RAG pipelines with document ingestion and vector search.
• Containerize services with Docker and deploy on Kubernetes.
• Instrument observability and evaluation harnesses using Langfuse.

Skills

Python development

Backend service design

ML/NLP expertise

Risk modeling domain knowledge

Education

Bachelor's degree in Computer Science, Engineering, or related field

Tools

LangChain

LangGraph

CrewAI

AutoGen

Semantic Kernel

Langfuse

FastAPI

Docker

Kubernetes

AWS/Azure/GCP

OpenSearch

Pinecone

Weaviate

Job description

Full-Stack Gen AI Lead Engineer – Model/Anlys/Valid Sr Analyst (C12)

The AI Lab is the engineering core of our Risk Modeling Solutions (RMS) team, focused on integrating Gen AI solutions into our risk management framework. The team is responsible for building and deploying practical, high‑impact applications that combine deep quantitative analysis with cutting‑edge AI. These solutions enhance analytical decision‑making, automate complex reporting, and create significant operational efficiencies for the business.

Responsibilities

• Architect Agentic Systems: Design and lead the implementation of complex, multi‑agent AI workflows capable of advanced reasoning, planning, and autonomous execution using frameworks like LangGraph, CrewAI, and Google ADK.
• Solution Design & Development: Translate complex business problems within the risk domain into well‑defined technical requirements, and develop robust, end‑to‑end AI solutions to address them.
• AI Workflow Development: Implement end‑to‑end agentic AI workflows using frameworks like LangGraph, CrewAI, and AutoGen, focusing on reasoning, tool use, and memory.
• LLM Orchestration: Build and optimize retrieval pipelines, memory layers, and tool‑use sequences using frameworks like LangChain.
• Backend & API Engineering: Develop robust, scalable Python‑based microservices and REST APIs using FastAPI to expose AI capabilities.
• RAG Implementation: Construct and refine Retrieval‑Augmented Generation (RAG) pipelines, including document ingestion, embedding, and vector search integration with databases like Azure AI Search or Pinecone.
• Containerization & Deployment: Package AI services using Docker and deploy them on Kubernetes, contributing to CI/CD pipelines for smooth and reliable releases.
• Observability & Evaluation: Instrument AI workflows using platforms like Langfuse for tracing and debugging. Implement and maintain evaluation harnesses to ensure model quality and performance.

Qualifications

• 7+ years of professional experience in a role blending software development and data science/machine learning.
• Expert‑level Python development skills and a proven track record of designing and building scalable backend services and APIs (FastAPI preferred).
• Deep, hands‑on experience designing and building solutions with multiple agentic frameworks (e.g., LangChain, LangGraph, CrewAI, AutoGen, Semantic Kernel).
• Extensive experience architecting and optimizing RAG systems and integrating with vector databases (e.g., OpenSearch, Pinecone, Weaviate).
• Proven expertise in designing and deploying containerized (Docker/Kubernetes) AI systems on a major cloud platform (AWS, Azure, or GCP).
• Strong experience implementing MLOps principles, including CI/CD, observability, and evaluation frameworks for LLM‑based systems.
• In‑depth understanding of AI risk, safety, and enterprise governance requirements.
• Strong background in ML, deep learning, and NLP, including Transformer architectures.

Preferred Qualifications

• Experience leading the design of LLM evaluation harnesses for automated release validation.
• Deep experience with Langfuse or similar AI observability and tracing platforms.
• Hands‑on expertise with AWS Bedrock, Azure AI Foundry, or GCP Vertex AI.
• Knowledge of GraphRAG patterns and advanced multi‑hop retrieval strategies.
• AWS certifications (e.g., Solutions Architect, AI/ML Specialty) or equivalent.
• Background in financial services or another highly regulated industry.

Education

Bachelor's degree in Computer Science, Engineering, Business, or a related field.

This is an Individual contributor role.

Citi is an equal opportunity employer, and qualified candidates will receive consideration without regard to their race, color, religion, sex, sexual orientation, gender identity, national origin, disability, status as a protected veteran, or any other characteristic protected by law.

If you are a person with a disability and need a reasonable accommodation to use our search tools and/or apply for a career opportunity, view Citi’s EEO Policy Statement and the Know Your Rights poster.