JPMorgan Chase—Director - Applied AI ML (Software Engineering/Data & Agentic Systems)

Job description

Description: Join our innovative team and shape the future of software development with AI-driven solutions. As an Applied GenAI Engineering Director within the Ai4Tect Team at JPMorganChase, you will leverage deep software engineering and data engineering expertise to design, build, and deliver trusted, secure, stable, and scalable GenAI capabilities. You will partner with agile engineering teams to create production-grade agentic workflows and RAG-based systems, and to establish reusable frameworks (“golden paths”), shared components, and best practices that accelerate delivery and ensure consistency across business functions. You will stay current on GenAI engineering trends and translate complex technical tradeoffs into clear guidance for senior stakeholders, enabling informed decisions and measurable business impact. Job Responsibilities - Establish and promote a library of reusable GenAI/ML engineering assets, including reference implementations, standardized templates/SDKs, shared RAG components (ingestion, chunking, embedding, indexing, retrieval), and deployment patterns. - Lead the creation of shared tools and platforms that streamline the end-to-end lifecycle for GenAI applications, including data pipelines, orchestration, evaluation, monitoring/telemetry, and release governance. - Build and operationalize agentic GenAI workflows (planning/execution patterns, tool calling, state management, retries) with appropriate guardrails, permissions, and observability. - Design and implement Generative AI evaluation and feedback loops (offline test suites, human review where needed, continuous evaluation, telemetry-based monitoring, regression gating in CI/CD). - Advise on strategy and development across multiple GenAI products, applications, and technology portfolios—focusing on common capabilities that scale across teams rather than one-off solutions. - Serve as a lead advisor on technical feasibility and business value for GenAI use cases, driving build-vs-buy decisions and pragmatic solution designs. - Liaise with firmwide AI/ML stakeholders to drive standards, interoperability, adoption, and reuse of shared frameworks. - Communicate complex technical issues and tradeoffs (quality vs latency vs cost; evaluation design; governance; security) to leadership to support well-informed strategic decisions. - Influence across business, product, and technology teams; effectively manage senior stakeholder relationships; mentor engineers and practitioners to raise engineering and delivery standards. - Champion the firm’s culture of diversity, opportunity, inclusion, and respect. Required qualifications, capabilities, and skills - 10+ years of applied experience in software engineering and/or data engineering using Python, Java, or similar languages, building production distributed systems end-to-end. - Hands-on experience designing and delivering GenAI systems to production, including RAG (embeddings, retrieval/indexing) and evaluation/monitoring. - Hands-on experience building agentic workflows (tool calling, orchestration, state, retries, guardrails) using frameworks such as LangChain/LangGraph or equivalent. - Strong understanding of data architecture and engineering (lakehouse/data platform concepts), including data quality, lineage/metadata, idempotent pipelines, backfills, and governance/PII controls relevant to GenAI. - Strong cloud-native experience on AWS, including secure deployment and operations (e.g., EKS and/or managed services), plus cost/latency management. - Proven ability to translate complex technical issues to senior stakeholders; excellent communication, attention to detail, and follow-through. Preferred qualifications, capabilities, and skills - Bachelor’s/Master’s degree in Computer Science (or equivalent practical experience). - Working knowledge of PyTorch or TensorFlow (enough to partner effectively with ML practitioners). - Experience with ML/GenAI evaluation automation and CI/CD quality gates (beyond basic offline testing).