Software Engineer III

Build products that help teams deliver real business outcomes with agent-based artificial intelligence-safely, reliably, and with measurable impact. You will join a high-visibility platform team focused on accelerating engineering delivery through reusable capabilities, strong engineering practices, and a culture of thoughtful collaboration and continuous learning.

As a Software Engineer III at JPMorganChase within Enterprise Technology, AI and Machine Learning and Data Platforms, you will design, build, and operate platform capabilities that enable teams to develop, evaluate, and run agent-based solutions powered by large language models (LLMs). You will translate complex problems into production-grade services, partnering with product, engineering, data, security, and risk and compliance stakeholders to deliver measurable outcomes. You will take end-to-end ownership-from design and implementation through monitoring and continuous improvement-while maintaining a high bar for quality, reliability, and responsible use.

Job responsibilities

• Deliver end-to-end platform features for agent-based artificial intelligence use cases, from technical design through production deployment, monitoring, and ongoing improvements
• Build and maintain reusable services and components that enable teams to create, evaluate, and operate LLM-powered agents at scale
• Implement evaluation and testing approaches for LLM systems, including quality measurement, regression testing, and error analysis to improve reliability over time
• Develop observability capabilities such as tracing, metrics, logs, and analytics to support healthy operations and fast troubleshooting
• Apply security, safety, and governance-by-design patterns, including guardrails, access controls, and audit-ready operational practices aligned to security and governance requirements
• Collaborate with product managers and stakeholders to define success metrics, prioritize work, and deliver measurable business impact
• Partner across engineering, data, security, and risk and compliance stakeholders to ensure solutions are secure, stable, and scalable
• Produce clear documentation and reference implementations that accelerate adoption and promote responsible, consistent engineering practices

Required qualifications, capabilities and skills

• Formal training or certification on software engineering concepts and 3+ years applied experience
• Proficiency in Python with strong software engineering fundamentals, including testing practices, version control, and code review
• Experience building and operating production services, including incident readiness, performance tuning, and operational stability for data-intensive systems
• Hands-on experience delivering machine learning and/or LLM-enabled features into production environments, including monitoring and post-deployment iteration
• Practical experience with prompt engineering and retrieval-augmented generation (RAG), including evaluation methods and quality measurement
• Ability to design maintainable systems and make sound technical decisions through ambiguity, balancing delivery speed, risk, and long-term sustainability
• Strong communication skills, including the ability to explain technical trade-offs to both technical and non-technical stakeholders
• Strong collaboration skills and demonstrated ability to partner effectively across teams to deliver outcomes

Preferred qualifications, capabilities and skills

• Experience with agent orchestration frameworks (for example, LangGraph, LlamaIndex, or comparable orchestration approaches) and LLM evaluation tooling
• Experience with continuous integration/continuous delivery (CI/CD) practices and production deployments using Docker and Kubernetes
• Familiarity with vector databases, embedding pipelines, and related approaches used in retrieval-augmented generation solutions
• Experience with cloud and machine learning platforms such as Amazon Web Services or Databricks (or comparable platforms)
• Experience contributing to governance, validation approaches, or guardrail frameworks for enterprise-scale artificial intelligence solutions