AI Engineer III - Agentic AI

About American Express Technology

Role Overview

Responsibilities

• Develop LLM-powered and agentic product features and workflows that reason over context, call tools, and perform actions.
• Implement and maintain RAG pipelines over financial data with emphasis on correctness and safety.
• Contribute to LLM services, orchestration components, evaluation tooling, and monitoring infrastructure.
• Monitor, debug, and improve the reliability and performance of AI systems in production.
• Collaborate with product and design partners to turn customer needs into technical solutions.
• Use AI-assisted development tools for design, implementation, testing, debugging, and refactoring.

Technical Environment

• Core technologies include Python, Go, TypeScript, AWS or GCP, Kubernetes, REST, gRPC, event-driven architectures, and Kafka.
• Agentic AI work includes commercial and open-source LLMs, orchestration, retrieval-augmented generation, vector storage, evaluation, and structured data handling.
• Systems are built for reliability, security, auditability, and compliance in a financial services environment.

Qualifications

• 4+ years of professional software engineering experience.
• Hands-on experience with AI-powered features, LLM-based applications, or applied machine learning systems.
• Engineering fundamentals in at least one backend language, including Python, Go, or TypeScript.
• Familiarity with APIs, basic cloud concepts, and modern development practices.
• Interest in agentic AI, autonomy, and AI-assisted development workflows.
• Willingness to learn, accept feedback, and grow technical ownership.
• Comfort working in collaborative, cross-functional teams and a strong customer mindset.

Preferred Qualifications

• Exposure to LLM tooling, prompt engineering, RAG, or agent frameworks through work, coursework, or personal projects.
• Internship or early-career experience in fintech or another regulated environment.
• Contributions to open-source projects, hackathons, or AI and developer-tooling side projects.

Workplace Model

• American Express states that roles may be onsite, hybrid, or fully virtual depending on the role and business needs; this posting does not specify the arrangement for this position.