Lead Software Engineer - Agentic AI

JPMorganChase is building agent-based capabilities for the Consumer and Community Banking Deposits 2.0 platform, and the team needs a lead engineer to design and drive delivery of an agentic AI platform. This role partners across product, engineering, risk, and control stakeholders to set direction, raise engineering standards, and help translate experimentation into production-ready architecture.

Working onsite in Plano, TX, you will lead end-to-end delivery of core agent platform components, establish quality and operational expectations for agent workloads, and embed responsible AI governance into platform design.

What you’ll do

• Execute creative software solutions across design, development, and technical troubleshooting, using non-routine approaches to break down complex problems.
• Define and drive the platform roadmap for agent-based capabilities with measurable outcomes, reliability, and usability.
• Lead end-to-end delivery of core agent platform components, including software development kits, reference implementations, and integration patterns.
• Partner with product, engineering, risk, and control stakeholders to align requirements, prioritize trade-offs, and unblock execution.
• Establish quality, performance, and operational standards for agent workloads, including monitoring, testing, and incident readiness.
• Translate experimentation into production by driving clear architecture decisions, scalable designs, and repeatable deployment practices.
• Guide responsible development practices by embedding governance, privacy, and model risk considerations into platform design.
• Communicate technical strategy and progress to senior stakeholders with clarity, data, and pragmatic recommendations.
• Drive team adoption of enterprise-authorized AI-assisted engineering practices, including AI-assisted code review or refactoring, test strategy acceleration, and incident or root-cause analysis support.
• Set consistent validation standards (secure coding, peer review, automated testing) and promote reuse of proven patterns across the team.
• Apply knowledge of Software Development Life Cycle toolchain capabilities, including enterprise-authorized AI-assisted development and automation, to improve the value realized by automation.

Requirements

• 5+ years applied experience with formal training or certification on software engineering concepts.
• Proficiency in Python (primary for agent orchestration and LLM tooling) and/or TypeScript, Java, or Go for enterprise backend integration.
• Data and RAG systems experience, including designing hybrid search pipelines (dense vector retrieval, BM25, rerankers) with vector databases such as Pinecone, Milvus, Qdrant, or pgvector.
• Backend and API design experience building scalable microservices using FastAPI, Spring Boot, or Node.js to expose agent interfaces via REST, WebSockets, and Server-Sent Events for streaming tokens and tool calls.
• Demonstrated experience leading effective use of approved AI-assisted software development tools for coding, code review, test acceleration, and troubleshooting, including setting team expectations for validating AI outputs for correctness, performance, and security.
• Strong understanding of responsible AI use in engineering workflows, including data sensitivity, secure handling of inputs and outputs, and resiliency and security expectations, with experience coaching engineers on safe, compliant adoption.
• Production experience with agentic frameworks and orchestration (examples include LangGraph, AutoGen, CrewAI, LlamaIndex Workflows, Semantic Kernel).
• Deep expertise in tool calling and function calling, including structuring model tool calls, JSON schema validation, dynamic API integration, sandboxed code execution, and MCP (Model Context Protocol).
• Architecture and memory management experience implementing short-term and episodic memory (scratchpads, state graphs, vector-based retrieval, conversational buffer compaction).
• LLM foundations including advanced prompt engineering, chain-of-thought, ReAct (Reasoning + Acting), reflection loops, and output grounding or guardrails (examples include NeMo Guardrails and Guardrails AI).

Technologies you’ll work with

• Languages and frameworks: Python, TypeScript, Java, Go, FastAPI, Spring Boot, Node.js, REST, WebSockets, Server-Sent Events
• Agent tooling: LangGraph, AutoGen, CrewAI, LlamaIndex Workflows, Semantic Kernel
• Data and retrieval: Pinecone, Milvus, Qdrant, pgvector, BM25, dense vector retrieval, rerankers
• Validation and protocols: JSON schema validation, MCP (Model Context Protocol)
• Guardrails: NeMo Guardrails, Guardrails AI

About the team

• The Consumer & Community Banking Group relies on innovators to serve consumers, small businesses, municipalities, and non-profits.
• You will help deliver tools and services spanning personal and small business banking, as well as lending, mortgages, credit cards, payments, auto finance, and investment advice.
• The group is focused on cutting-edge mobile applications, digital experiences, and next generation banking technology solutions.