Applied Ai/ml Lead - Vice President - Payments

JPMorgan Chase JPMorgan Chase · Banking · New York, NY +1 · Commercial & Investment Bank

Lead Applied AI/ML role focused on delivering end-to-end ML capabilities for payments, including fraud detection, operational automation, and client experience. Responsibilities include designing, building, and deploying production-grade ML systems, defining evaluation metrics, establishing model lifecycle standards, and partnering with risk/compliance teams. Requires strong hands-on expertise in neural networks, Transformers, fine-tuning, distillation, and software engineering skills in Python with PyTorch/TensorFlow.

What you'd actually do

  1. Own a major AI/ML problem area end-to-end (e.g., transaction risk & fraud, anomaly detection, payment exceptions automation, routing/authorization optimization), from opportunity sizing and problem framing through production rollout and iteration.
  2. Design and build production-grade ML systems that operate at payments scale, balancing accuracy, latency, throughput, and cost across batch, near-real-time, and real-time inference patterns.
  3. Develop state-of-the-art neural approaches including Transformer architectures, representation learning, and sequence/graph methods where appropriate; apply fine-tuning (full/parameter-efficient), distillation, and model compression techniques to meet deployment constraints.
  4. Define rigorous evaluation and measurement: offline metrics, calibration, robustness testing, segmentation, and online experimentation where feasible; translate model lift into business impact (loss reduction, approval rates, false-positive reduction, ops productivity, client outcomes).
  5. Establish model lifecycle standards: reproducibility, testing, monitoring/alerting, drift detection, champion–challenger approaches, incident response/rollback, and ongoing performance governance.

Skills

Required

  • Master's or PhD in Computer Science, Machine Learning, Statistics, Mathematics, Operations Research, or related field, plus 6+ years of industry experience delivering applied ML (or PhD with equivalent applied experience).
  • Demonstrated track record of shipping ML models into production with measurable business impact, including iteration post-launch (monitoring, retraining, recalibration, and continuous improvement).
  • Strong hands-on expertise in neural networks and Transformers, including practical experience with: Fine-tuning strategies (e.g., full fine-tuning and parameter-efficient methods), Distillation / compression (teacher–student, quantization-aware approaches, latency/cost-driven optimization), Robust evaluation and failure-mode analysis for real-world deployment
  • Strong software engineering skills in Python, with deep experience in PyTorch or TensorFlow and standard ML/data libraries.
  • Experience building data-driven systems using SQL and distributed processing (e.g., Spark/PySpark or equivalent).
  • Cloud and production experience on AWS (or equivalent cloud), including deploying services/pipelines and operating them reliably at scale.
  • Ability to take ambiguous business problems and turn them into structured ML plans: data strategy, modeling approach, evaluation, rollout, and operationalization.
  • Excellent communication skills, including explaining tradeoffs (accuracy vs latency, risk vs customer friction, complexity vs maintainability) to both technical and non-technical audiences.

Nice to have

  • Payments domain experience: fraud/risk, transaction monitoring, identity/account takeover, disputes/chargebacks, sanctions/AML-adjacent signal work, payment exceptions and investigations, routing/authorization optimization, or treasury/transaction banking.
  • Experience with streaming/real-time architectures and feature generation (e.g., event-driven systems, point-in-time correctness, leakage prevention).
  • Strong ML platform/MLOps exposure: feature stores, model registries, CI/CD for ML, scalable training/inference, observability, and governance workflows.
  • Experience with Docker/Kubernetes and modern data platforms (e.g., Databricks, Snowflake) where relevant.

What the JD emphasized

  • shipping ML models into production with measurable business impact
  • iteration post-launch (monitoring, retraining, recalibration, and continuous improvement)
  • Fine-tuning strategies (e.g., full fine-tuning and parameter-efficient methods)
  • Distillation / compression (teacher–student, quantization-aware approaches, latency/cost-driven optimization)
  • Robust evaluation and failure-mode analysis for real-world deployment

Other signals

  • end-to-end delivery of high-impact AI/ML capabilities
  • build, fine-tune, distill, and deploy models
  • improve fraud/risk outcomes, operational automation, and client experience
  • set technical direction, establish standards, and lead execution through influence
  • partnering closely with Product, Engineering, Risk, Compliance, and Data teams
  • Design and build production-grade ML systems that operate at payments scale
  • Define rigorous evaluation and measurement
  • Establish model lifecycle standards
  • Partner with Risk/Compliance and model governance stakeholders
  • Lead through influence
  • Communicate clearly to senior stakeholders