New opportunity

Applied AI Technical Product Manager

BNY · New York, United States

About this role

BNY seeks an Applied AI Technical Product Manager for its Innovation Team to lead AI-driven business process transformation across the bank. The role combines product strategy with technical leadership to build scalable, secure and explainable agentic and intelligent automation solutions that improve efficiency, reduce risk and enhance client experiences. It is based in New York, with four days in the office each week.

Responsibilities

Own the vision, strategy and roadmap for AI-powered process transformation, aligning priorities with enterprise modernization and operational goals. Identify opportunities for AI, LLMs and agentic architectures to replace manual workflows and improve decision-making; turn complex business problems into product principles, PR/FAQs and execution plans.
Lead cross-functional product delivery with Engineering, Data Science, Operations, Risk and Compliance. Oversee solutions involving event-driven architecture, APIs, microservices, workflow engines, state management, metadata-aware retrieval and multi-agent coordination. Support production deployment through model lifecycle management, observability, guardrails, performance optimization and reliability standards.
Ensure responsible AI governance covering bias, explainability, privacy, security and operational resilience. Lead a product team, resolve delivery blockers, communicate with senior executives and drive adoption through feedback mechanisms, playbooks and change management.
Track productivity and efficiency gains, including examples of 20–50% time reduction, alongside cost savings, automation impact, risk and compliance improvements, and client experience.

Required: A bachelor's degree in STEM, Economics or a related discipline; approximately ten years of progressive experience in technical product management, program leadership or AI-driven platform delivery; knowledge of generative AI, LLM systems, agentic orchestration, distributed systems, APIs, retrieval systems and AI performance metrics; experience deploying enterprise AI with tenant-aware access controls, observability and production reliability; and the ability to influence senior leaders.

Preferred

Experience in regulated industries or deploying AI for Big Tech or Fortune 500 enterprises; an advanced degree; or CSPO, A-CSPO, SAFe POPM, CSM, AWS Cloud Practitioner or Generative AI specialization certification.

The expected starting base salary is $127,000–$250,000 per year, determined individually. Depending on the position, total compensation may include incentives and benefits. BNY also offers wellbeing programs and paid leave, including paid volunteer time.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Technical product managementProduct strategyProduct roadmappingGenerative AILarge language modelsAI agentsIntelligent automationAI orchestrationDistributed systemsMicroservicesEvent-driven architectureAPIsWorkflow enginesState managementMetadata indexingInformation retrievalMulti-agent coordinationAI lifecycle managementModel performance metricsObservabilityTelemetryAccess controlsResponsible AIExplainabilityData privacyData securityStakeholder managementChange managementSystems thinkingCSPO certification readiness

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Set the AI transformation roadmap
The Director, Technical Product Manager owns the vision and roadmap for AI-driven process transformation, aligned with enterprise modernization and operational efficiency. The role calls for identifying high-value manual workflows and turning ambiguous problems into product tenets, PR/FAQs, and execution plans.
Scale production-ready AI workflows
The role oversees solutions spanning event-driven architecture, APIs, workflow engines, metadata-aware retrieval, and multi-agent coordination. It also partners with Engineering and Data Science on model lifecycle management, observability, guardrails, performance, and reliability.
Build responsible AI into delivery
The posting makes bias, explainability, privacy, data security, and operational resilience explicit responsibilities. It also requires experience operationalizing enterprise AI with secure tenant-aware access controls.
Lead alignment and adoption
This director leads a product team and works across Product, Engineering, Data Science, Operations, Risk, and Compliance. Influencing senior executives, removing blockers, using field feedback, and supporting adoption through playbooks and change management are part of the remit.
Prove business impact
The role defines and tracks productivity, cost, automation, risk, compliance, and client-experience outcomes. The posting gives 20–50% time reduction as an example metric, not a guaranteed result for every initiative.

Likely challenges to think through

Choosing where agents add value
Preparation hypothesis: an AI-first roadmap may need to distinguish workflows suited to agentic systems from those better served by simpler automation, while avoiding complexity without measurable benefit. Prepare a prioritization approach that weighs workflow value, decision quality, risk, and a baseline for measuring improvement.
Moving quickly without weakening controls
Preparation hypothesis: experimentation and delivery speed may compete with explainability, access controls, and production reliability in a regulated enterprise. Prepare to explain how you would stage releases, evaluate model and workflow behavior, monitor failures, and establish guardrails with Risk and Compliance.
Turning executive support into sustained use
Preparation hypothesis: investment approval alone may not change established workflows across enterprise functions. Prepare an example or approach showing how to reconcile stakeholder priorities, communicate outcomes to executives, gather field feedback, and adjust playbooks or the roadmap when adoption lags.

Practice your answers

Technical practice

How would you prioritize a roadmap for AI-driven process transformation across enterprise functions?

What to coverSuggested answer focus: Explain how you would weigh operational efficiency, risk reduction, client experience and alignment with enterprise modernization. Describe how you would make trade-offs visible to business stakeholders and revise priorities as evidence emerges.

How would you decide whether a manual workflow calls for an LLM, agentic automation or a simpler approach?

What to coverSuggested answer focus: Compare the workflow’s decision complexity, need for judgment, data availability and cost of errors. Explain when added AI capability would justify its governance and reliability burden.

How would you turn an ambiguous business problem into product tenets, a PR/FAQ and an execution plan?

What to coverSuggested answer focus: Start with the user problem and desired outcome, then define scope, assumptions, success measures and decisions requiring stakeholder agreement. Show how you would test uncertain assumptions before committing to a scalable plan.

How would you structure an AI-enabled workflow using events, APIs, microservices and a workflow engine?

What to coverSuggested answer focus: Explain the responsibilities of each component, how state moves through the workflow and where integration boundaries belong. Discuss trade-offs between loosely coupled services and the operational complexity of distributed delivery.

How would you design metadata-aware retrieval with secure tenant-aware access?

What to coverSuggested answer focus: Describe how metadata could support relevant retrieval while access controls restrict what each tenant can retrieve. Consider how you would evaluate relevance and investigate a retrieval result that is wrong or improperly scoped.

When would multi-agent coordination improve a process, and how would you control it?

What to coverSuggested answer focus: Identify a reason to divide work among agents rather than use a simpler workflow. Discuss coordination, state management, guardrails and how the system would handle a failed or uncertain step.

How would you manage the lifecycle and performance of an LLM-powered product?

What to coverSuggested answer focus: Outline evaluation before release, monitoring after deployment and a process for responding to degraded results. Distinguish model-quality measures from the operational and business outcomes the posting asks this role to track.

How would you troubleshoot a production AI workflow that has become slow or unreliable?

What to coverSuggested answer focus: Use observability and telemetry to locate the affected step, assess its impact and separate model, retrieval and integration issues. Explain how you would balance a prompt recovery with guardrails and production reliability.

How would you evaluate responsible-AI risks before scaling an automation product?

What to coverSuggested answer focus: Address bias, explainability, privacy, data security and operational resilience, all named in the posting. Explain how Product, Risk and Compliance could use evidence about those risks to inform a release decision.

How would you demonstrate that an AI workflow delivers measurable business value?

What to coverSuggested answer focus: Establish a baseline and define measures for time saved, cost, automation impact, risk or compliance, and client experience. Explain how you would avoid attributing every observed improvement to the product; the posting’s 20–50% time reduction is an example, not a promised result.

Behavioural practice

Tell me about a time you influenced senior leaders on a difficult product investment decision.

What to coverSuggested answer focus: Use situation, action and result to show how you framed options, communicated technical trade-offs in business terms and reached a decision. Be clear about your own contribution and the outcome you could substantiate.

Describe a time Product, Engineering and risk-focused stakeholders disagreed about a release.

What to coverSuggested answer focus: Use situation, action and result to explain the disagreement, how you surfaced the underlying risks and how the group resolved or escalated them. Include the effect on delivery and governance without implying that speed always prevailed.

Tell me about a time you removed a blocker across concurrent workstreams.

What to coverSuggested answer focus: Use situation, action and result to describe the dependencies, how you aligned owners and priorities, and what changed. Highlight how you maintained delivery visibility while leading the team.

Describe a time you drove adoption of a significant workflow or product change.

What to coverSuggested answer focus: Use situation, action and result to explain the resistance or uncertainty, the communication or playbooks you introduced, and how you checked adoption. Distinguish rollout activity from evidence that the change improved work.

Tell me about a time frontline feedback changed your product direction.

What to coverSuggested answer focus: Use situation, action and result to show how feedback reached the product team, how you weighed it against the roadmap and what you changed. Include a measurable result or a clear account of what you learned.

Sharpen your knowledge

Topics to prepare

AI product strategy and prioritization
The posting calls for a vision and roadmap aligned to modernization and operational efficiency. Revision exercise: compare two hypothetical manual workflows and explain which to prioritize, using expected value, delivery complexity and risk. The comparison is a preparation exercise, not a stated BNY process.
LLM, agentic and retrieval design
The required knowledge includes LLM systems, agentic orchestration, metadata indexing and retrieval. Revision exercise: explain when a workflow needs retrieval or multiple agents, and when simpler automation would suffice; identify how retrieved information would be checked before use.
Enterprise integration architecture
The role oversees event-driven design, APIs, microservices, workflow engines and state management. Revision exercise: trace a hypothetical request through these components and explain how state is preserved if a step fails.
AI lifecycle and production reliability
The posting requires experience with observability, telemetry and production reliability, and calls for model lifecycle management, guardrails and performance optimization. Revision exercise: outline what to monitor after release, what would trigger investigation and how a change would be evaluated before wider use.
Responsible AI and access governance
BNY explicitly names bias, explainability, privacy, security, operational resilience and secure tenant-aware access controls. Revision exercise: review a hypothetical AI workflow for each risk and explain what evidence a product leader would seek before launch.
Business outcomes and adoption
The role tracks efficiency, cost, automation, risk, compliance and client-experience outcomes while driving enterprise adoption. Revision exercise: distinguish an activity metric from an outcome metric and explain how field feedback could change roadmap priorities. The posting’s 20–50% time reduction is an example, not a promised result.

Skills to revise

Draft a product tenet and PR/FAQ
Practice turning one hypothetical manual workflow into a product tenet, PR/FAQ and phased roadmap. This directly reflects the posting’s requirement to translate ambiguous problems into execution plans; the workflow is only an interview exercise.
Sketch an end-to-end workflow
Draw a simple design showing an event, API integration, workflow state, metadata-aware retrieval and any agent handoffs. Annotate an access-control check and a failure path. This practices the technical oversight described in the role, without assuming a particular BNY implementation.
Build an impact measurement plan
Create a small metric table with a baseline, measurement period and proposed measures for time saved, cost, automation impact and risk. Explain how you would avoid attributing every change to AI. These outcome categories are stated in the posting; the measurement design is a suggested exercise.
Prepare a production-readiness review
Draft questions for Engineering, Data Science, Risk and Compliance on model performance, telemetry, guardrails, tenant-aware access and recovery from failures. This rehearses the cross-functional production and governance responsibilities explicitly described.
Present an executive decision brief
Practice a short investment-prioritization narrative covering the problem, proposed roadmap, expected outcomes, principal risks and an adoption feedback loop. The posting specifically calls for senior-leadership influence, executive storytelling and structured change management.

Your preparation checklist

Assemble a role evidence sheet
The posting calls for a relevant bachelor’s degree, roughly ten years of progressive experience, enterprise AI delivery, and senior-leader influence. Prepare a one-page list of applicable examples and evidence; mark any gaps rather than inventing experience.
Rehearse an AI platform walkthrough
If you have a relevant project, prepare a concise account of the business problem, your product decisions, cross-functional delivery, production rollout, and measured results. Distinguish your contribution from the team’s.
Draft a process-transformation proposal
As interview preparation, choose a hypothetical manual enterprise workflow and write product tenets, a brief PR/FAQ, and a prioritized roadmap. Explain where AI adds value and where a simpler approach would suffice.
Sketch a production architecture
Prepare a diagram showing how an agentic workflow could use events, APIs, workflow state, metadata-aware retrieval, and orchestration. Be ready to explain trade-offs, failure handling, and when multi-agent coordination is warranted.
Build a governance and operations checklist
The role emphasizes tenant-aware access, observability, reliability, and responsible AI. Prepare review questions covering permissions, privacy, bias, explainability, guardrails, model monitoring, and operational resilience.
Prepare an impact and adoption narrative
Outline how you would establish baselines and track time saved, cost, automation impact, risk, compliance, and client experience without assuming the posting’s example of 20–50% time reduction is guaranteed. Include how feedback and playbooks would support adoption.
Bring executive-level questions
Prepare questions about which enterprise workflows are priorities, how Product partners with Engineering, Operations, Risk, and Compliance, what production standards apply, and how leadership judges value across concurrent workstreams.

Availability last checked: Nov 4, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…