AI and Data Product Manager

Job Purpose

Intercontinental Exchange, Inc. (ICE) is seeking an AI and Data Product Manager to lead the transformation of established business workflows by rebuilding them around AI-not by layering AI on top of them. This role owns the strategy, roadmap, and delivery of our AI-powered applications and data products, sitting at the intersection of business operations, data science, machine learning engineering, and our customers. You will be part of a highly visible team central to ICE’s strategy to analyze mortgage and market data and deliver AI-driven insights to our clients in a meaningful, responsible, and scalable way.

This is a process-transformation role first and a feature-delivery role second. We hire for the ability to decompose workflows and apply AI where it is verifiable-not for prior expertise in any specific industry. A track record of entering an unfamiliar domain, mapping its workflows, and shipping something measurable is the signal we value most.

Responsibilities

• Map and decompose existing business workflows end-to-end-identifying steps that are high-volume, high-variance, and verifiable-before deciding where AI belongs.

• Reimagine processes around AI rather than bolting AI onto current steps, prioritizing opportunities by the principle of volume, variance, and verifiability.

• Define and own the product vision, strategy, and multi-quarter roadmap for a portfolio of AI applications and data products aligned to business objectives.

• Size AI opportunities with conservative, evidence-based ROI assumptions, targeting tasks whose outputs can be reliably graded and avoiding “too much, too fast” over-commitment.

• Partner closely with data science and ML engineering to translate models-including predictive analytics, NLP, and LLM/generative AI and agentic solutions-into reliable, production-grade products.

• Design evaluation criteria and acceptance thresholds (“define good before building”); establish evals, blind review panels, and LLM-as-judge methods, and monitor for hallucination, bias drift, and model degradation in production.

• Architect human-in-the-loop workflows with expert review designed in, expanding automation only after each phase is proven.

• Productize ICE’s proprietary data assets into well-defined data products such as APIs, data feeds, datasets, dashboards, and embedded analytics.

• Write clear product requirement documents (PRDs), user stories, and acceptance criteria; maintain and prioritize the product backlog within an Agile/Scrum environment.

• Define success metrics and KPIs (adoption, task success rate, model performance, revenue, ROI) and use data to measure outcomes and continuously improve products.

• Drive change management and adoption-bridging data scientists and business owners and getting non-technical stakeholders to embrace AI-changed workflows.

• Champion responsible AI in partnership with data science, risk, and compliance: model governance, bias and fairness, explainability, model risk, and data quality.

• Ensure products meet regulatory and data-governance requirements relevant to mortgage and financial services (e.g., MISMO, FNMA, FHLMC, GNMA, and applicable privacy standards).

• Communicate roadmap, trade-offs, progress, and results to cross-functional partners and executive leadership.

Knowledge and Experience

• 6+ years of product management experience, with demonstrated work building AI/ML-powered products, data products, or workflow-automation solutions (mid-level is the target tier).

• A demonstrable example of entering a domain cold, mapping its workflows, identifying AI leverage points, and shipping something measurable-industry independent.

• Strong process-decomposition skills: the ability to map a workflow in detail and score steps by volume, variance, and verifiability.

• Practical AI literacy: working comprehension of LLMs, RAG, agents, prompt engineering, and evaluation design (you do not need to code or train models).

• Empirical mindset: experience designing evals, blind reviews, A/B tests, and acceptance criteria, and iterating against evidence.

• Data literacy, including comfort with SQL and analytics tools to define metrics and inform decisions.

• Change-management and stakeholder-translation experience getting non-technical teams to adopt new, AI-driven ways of working.

• Ability to recall specific metrics from products you have shipped (e.g., hallucination rate, retrieval precision, task success rate, latency).

• Proven experience working in Agile/Scrum teams and managing a product backlog.

• Excellent written and oral communication, with the ability to explain probabilistic systems to both technical and non-technical audiences.

Preferred Knowledge and Experience

• Advanced degree (e.g., MBA) or product/Agile certification (e.g., Pragmatic Institute, CSPO, SAFe POPM).

• Hands-on experience with at least one workflow or process platform-e.g., n8n, Zapier, Make, Workato, Celonis, UiPath.

• Experience launching generative AI / LLM-based or agentic products or features.

• Background that develops process thinking before AI-operations management, management consulting, analytics/data, or growth/experimentation product management.

• Familiarity with cloud platforms (e.g., AWS) and modern data warehouses such as Snowflake or Databricks.

• Understanding of human-in-the-loop design, model monitoring, drift detection, and responsible-AI frameworks.

• Exposure to mortgage technology, capital markets, or financial services is helpful but not required.

• A computer science degree. Roughly 60% of working AI PMs do not hold one; demonstrated AI experience is the signal.

• The ability to code or train models. Data literacy and AI comprehension are sufficient.

• Prior expertise in mortgage or financial services. Pattern transfer and rapid domain immersion matter more than industry credentials; domain knowledge can be borrowed empirically from practitioners.

Technical & Tool Familiarity

• AI/ML Concepts: LLMs, RAG, agents, prompt engineering, and evaluation metrics (precision, recall, F1, hallucination rate, task success rate, latency).

• AI Orchestration (Execution): UiPath Maestro, n8n; awareness of agent protocols such as MCP, A2A, and ACP.

• Workflow Builders (Prototyping): Any of n8n, Zapier, Make, Workato.

• Data Platforms: SQL, Snowflake, Databricks, Spark; data pipelines and ETL concepts.

• Cloud: AWS (or comparable cloud environments).

• Visualization & BI: SIGMA, Tableau, Microsoft Power BI.

• Product & Delivery: Jira, Confluence, Productboard; product analytics.