Lead Machine Learning Engineer - Generative AI and Agent Platforms

You'll build and ship agents that real businesses depend on, not demos.

Platform already runs a federated portfolio of production agents - forecasting, anomaly detection, log analysis, with sales fulfillment and voice-of-client close behind - and you'll add to it.

You'll work across Corporate & Investment Bank sub-lines of business and Payments, using Platform's runtime, retrieval, and memory primitives plus platforms such as Databricks and the GenAI Gateway, and apply MLOps for automation, continuous delivery, and compliance with AI/ML control expectations.

Platform is the firm's agent runtime: it gives agents secure execution, agent-to-agent (A2A) communication, MCP-based tool access, a managed memory layer, and permission-aware, auditable operation in a regulated environment.

Your job is to turn that platform into shipped agents.

As an Applied AI and Machine Learning Lead at JPMorganChase within Payments Technology, Data Analytics, Regulatory & Compliance (Data Analytics) in the Commercial & Investment Bank, you will build and operate unique Agentic operation system and runtime that allows engineering teams build and safely, reliably run their AI agents.

You will also build own AI Agents that solve high-impact business problems with measurable outcomes.

You will own agent solutions end-to-end, combining strong engineering discipline with rigorous evaluation, safety, and operational excellence.

You will partner with product, engineering, and business stakeholders to deliver reliable, auditable AI capabilities at enterprise scale.

Job responsibilities


* Design and deliver production AI agents across a federated portfolio, owning solutions from prototype through launch and operational support


* Engineer retrieval systems that perform reliably in production, including hybrid retrieval-augmented generation patterns (vector search plus graph-based retrieval where appropriate) with robust chunking, ranking, and grounding approaches


* Implement agent memory patterns, including episodic and semantic memory with recall, summarization, and decay policies aligned to use-case needs


* Build entitlement-aware and tenant-aware context assembly so agents reason only over permitted data, supporting traceability and auditability


* Orchestrate multi-agent workflows and integrate external tools and data sources through secure connectors and standardized tool interfaces


* Develop evaluation frameworks, including task-level and end-to-end evaluations, regression suites, automated scoring, and release gates for quality and safety


* Deploy and operate agent services on public cloud platforms (Amazon Web Services and/or Microsoft Azure), applying strong software development lifecycle, security, resiliency, and observability practices


* Optimize runtime performance and reliability by instrumenting tracing, monitoring, and incident-response playbooks for agent services


* Partner with product and business leaders to transl...