New opportunity

Senior AI Infrastructure Platform Engineer, Assistant Vice President

State Street · Quincy, United States of America

About this role

State Street seeks a Senior AI Infrastructure Platform Engineer, Assistant Vice President, to help its Global AI Team design and deploy automation and AI solutions across business processes. The role combines infrastructure and platform architecture with hands-on full-stack development and leadership of generative and agentic AI projects.

Responsibilities

Lead development projects with business, IT and infrastructure teams, collaborating with data scientists, machine learning engineers, software engineers and QA engineers.
Roll out self-service technology platforms and build MVPs to assess feasibility and inform automation, AI and ML strategy. Design cloud-native applications and interoperable APIs for enterprise-scale use.
Review solution designs with Enterprise Architecture against its standards; support sprint planning and produce design, development and Jira artifacts.
Understand existing business processes and technology, distinguish standard from bespoke needs, and design future-state automation. Manage risks and take ownership of version control, model lifecycle and quality-control work products.

Qualifications and skills:

Bachelor’s degree in Computer Science or Engineering and 5+ years of technology development or consulting experience on large-scale implementations in consulting or financial services.
Hands-on expertise in Python, FastAPI, Unix, Docker, Postgres, LangChain, LangGraph, Kubernetes, Databricks, API management, AWS and Azure. Experience developing generative and agentic AI solutions, including model fine-tuning, selection, evaluation and inference.
Ability to design and deploy applications using Azure Foundry, OpenAI, Document Intelligence, Azure Functions, Azure Logic Apps, AWS Bedrock, SageMaker, AWS AgentCore, Step Functions, Lambda and EKS. Familiarity with Agile development, CI/CD, Git, Maven and Harness; strong system design, communication, analytical and problem-solving skills.
Preferred: cloud, AI or ML certification. Strong organization, ability to meet deadlines and experience with large cross-functional teams are a plus.

The listed locations are Quincy, Massachusetts; Princeton, New Jersey; Irvine, California; and Austin, Texas. The annual salary range is $90,000–$157,500 for the primary location and may differ elsewhere. Benefits include a 401(k) with company match, insurance, paid time off, an employee assistance program and eligibility for performance-based awards.

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

Skills for this role

PythonFastAPIUnixDockerPostgreSQLLangChainLangGraphKubernetesDatabricksAPI managementAWSAzureAzure AI FoundryOpenAIAzure AI Document IntelligenceAzure FunctionsAzure Logic AppsAmazon BedrockAmazon SageMakerAWS AgentCoreAWS Step FunctionsAWS LambdaAmazon EKSGenerative AIAI agentsModel fine-tuningModel selectionModel evaluationModel inferenceSystem architecture design and development principles and practices for AI and L

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

Lead while building
The role calls for independently leading GenAI and agentic AI projects while contributing hands-on full-stack development and architecture work alongside business, IT, infrastructure, data science, engineering and QA teams.
Deploy reusable AI platforms
The posting emphasizes rolling out self-service platforms and building cloud-native applications and interoperable APIs for enterprise-scale use.
Use MVPs to shape automation
MVPs are meant to test feasibility and inform automation, AI and ML strategy. The role also requires understanding existing processes and technology, separating standard from bespoke needs, and identifying client value.
Turn designs into execution
Responsibilities include presenting solutions for review against Enterprise Architecture standards, supporting sprint delivery, and producing design, development and Jira artifacts.
Own model and delivery quality
The role explicitly assigns proactive ownership of risks, version control, model lifecycle and quality-control work products; the posting also calls for model selection, evaluation and inference experience.

Likely challenges to think through

MVP speed versus enterprise readiness
A likely trade-off is testing an idea quickly without treating a prototype as deployment-ready. Prepare to explain how you would define feasibility evidence, then address architecture review, quality controls and lifecycle ownership before rollout.
Reusable platform versus bespoke needs
Self-service platforms must serve multiple processes, while the posting asks the engineer to distinguish standard from bespoke requirements. Prepare an approach for finding a reusable core, identifying justified exceptions and making the client-value trade-off explicit.
Interoperability versus platform complexity
The posting spans AWS, Azure, cloud-native applications and enterprise APIs. A plausible difficulty is preserving interoperability without adding unnecessary services or operational burden. Prepare to discuss service-selection criteria, API boundaries and deployment trade-offs rather than assuming every listed tool belongs in one solution.

Practice your answers

Technical practice

How would you design an enterprise-scale, cloud-native AI application?

What to coverPreparation suggestion: Explain the boundaries between the application, interoperable APIs, model services and infrastructure, and how you would assess scalability and operational risk. The posting names AWS and Azure; discuss how you would choose services for a specific use case rather than assuming both clouds must be used.

How would you build an interoperable AI API using Python and FastAPI?

What to coverPreparation suggestion: Cover request and response contracts, validation, error handling, and where Postgres would fit if persistent data were needed. Explain how you would test the API and manage changes for enterprise consumers.

When would you use an agentic workflow rather than a simpler GenAI application?

What to coverPreparation suggestion: Describe how you would structure steps and tool interactions with a framework such as LangGraph or LangChain, and how you would constrain and test the workflow. Address the trade-off between added capability and added complexity.

How would you select a model and decide whether to fine-tune it?

What to coverPreparation suggestion: Compare candidate models against the task, evaluation results, inference needs and implementation effort. Explain what evidence would justify fine-tuning instead of using an existing model.

How would you evaluate and control quality across a model’s lifecycle?

What to coverPreparation suggestion: Describe evaluation criteria, repeatable tests and review points before and after deployment. Connect model versioning and quality-control work products to the risk ownership stated in the posting.

How would you troubleshoot a containerized AI service that fails after deployment?

What to coverPreparation suggestion: Walk through a diagnostic sequence spanning the application, Docker image, configuration and Kubernetes deployment; the posting also names EKS. Explain how you would isolate the cause, verify a fix and avoid an uncontrolled change.

How would you choose between serverless components and a Kubernetes deployment?

What to coverPreparation suggestion: Use the listed Azure Functions, AWS Lambda, Step Functions and EKS as options, not as a required architecture. Compare workload shape, orchestration needs, operational ownership and scaling trade-offs.

How would you use an MVP to assess an automation opportunity?

What to coverPreparation suggestion: Start with the existing business process and technology, then distinguish standard needs from bespoke ones. Define what the MVP must demonstrate to inform feasibility and a future-state design, without assuming a particular business process.

What would you consider when rolling out a self-service AI platform?

What to coverPreparation suggestion: Explain how users would access reusable capabilities and APIs, and how you would balance ease of adoption with controlled releases and quality checks. Describe how feedback from business and technical teams would inform the rollout.

How would you take a solution design through architecture review and delivery?

What to coverPreparation suggestion: Describe the design decisions and artifacts you would bring to an Enterprise Architecture review, then how you would translate approved work into Jira stories. Connect Git-based version control, CI/CD and testing to traceable changes; the posting names these practices but does not prescribe a specific pipeline.

Behavioural practice

Tell me about a time you led a cross-functional development effort.

What to coverPreparation suggestion: Use a real situation involving different technical or business perspectives, then explain your own actions and the result. Make clear how you coordinated decisions while still contributing to the work yourself.

Describe a time you resolved disagreement about a solution design.

What to coverPreparation suggestion: Give the situation, the differing concerns, your actions to examine trade-offs, and the outcome. A design review or standards-related example would be relevant to the posting, but do not claim it occurred at State Street.

Tell me about a time you discovered that a requested solution did not fit the underlying business process.

What to coverPreparation suggestion: Explain the initial request, how you investigated the process and existing technology, what you changed, and the result. Highlight how you distinguished reusable needs from genuinely bespoke requirements.

Describe a time you managed a delivery risk while facing a deadline.

What to coverPreparation suggestion: State the risk and time constraint, the actions you took with the team or project manager, and the result. Keep the example tied to concrete delivery work, such as sprint scope, testing or release readiness.

Tell me about a time you took ownership of a quality issue.

What to coverPreparation suggestion: Give a real situation, your investigation and corrective actions, and the measurable or observable result. An example involving version control, a model lifecycle or a development work product would align with the role.

Sharpen your knowledge

Topics to prepare

Enterprise AI platform architecture
Suggested exercise: sketch a self-service AI platform with interoperable APIs, deployment boundaries and a path from MVP to enterprise use; explain what an Enterprise Architecture review would need to assess. The posting calls for platform rollout, cloud-native design and solution reviews against EA standards.
GenAI and agentic model lifecycle
Suggested exercise: compare model selection, fine-tuning, evaluation and inference choices for a hypothetical agentic application, including how quality would be checked over time. The posting explicitly names these lifecycle activities and ownership of quality-control work products.
AWS and Azure service choices
Suggested exercise: map the listed Azure services—Foundry, OpenAI, Document Intelligence, Functions and Logic Apps—against the listed AWS services—Bedrock, SageMaker, AgentCore, Step Functions, Lambda and EKS—for two possible application designs. Be ready to explain trade-offs rather than assume the role uses every service in one solution.
Business-process automation design
Suggested exercise: take a hypothetical business process, separate standard from bespoke needs, propose a future state and define what an MVP would test. The posting asks for process discovery, value assessment and MVPs to inform automation strategy.
Technical delivery and risk ownership
Suggested exercise: outline how a development lead would coordinate business, infrastructure, data science, engineering and QA contributors, then identify design artifacts, Jira stories, sprint decisions and risks. These responsibilities are stated in the posting.

Skills to revise

Build an interoperable Python API
Suggested exercise: implement a small Python and FastAPI service that reads and writes to Postgres; document its API contract and test a failure case. The posting specifies hands-on Python, FastAPI and Postgres expertise and API development for enterprise-scale cloud solutions.
Prototype an agentic workflow
Suggested exercise: use LangChain or LangGraph to build a small agentic flow, then record evaluation cases and explain the model-selection and inference choices. The posting names both frameworks and hands-on GenAI and agentic development.
Package and deploy a cloud-native service
Suggested exercise: containerize a small service with Docker, run it using Unix commands and prepare a Kubernetes deployment configuration; explain what would change for EKS. The posting names Docker, Unix, Kubernetes and EKS and calls for cloud-native deployment.
Implement a cloud orchestration path
Suggested exercise: design and, where practical, prototype one flow using Azure Functions and Logic Apps or AWS Lambda and Step Functions, with an explicit error path. These services are listed for application design and deployment; the particular flow is a preparation suggestion, not a stated project.
Practice version control and delivery artifacts
Suggested exercise: put a small change in Git, describe a CI/CD check, and write a Jira story plus a brief design and quality-control note for it. The posting names Git, CI/CD, Jira artifacts, version control and quality-control ownership.
Explain a Databricks integration
Suggested exercise: diagram how a Databricks-based data step could connect to a FastAPI application through an interoperable API, and identify the interface and failure points. Databricks and API management are listed as hands-on skills; this integration scenario is a suggested exercise, not a specified implementation.

Your preparation checklist

Map the role to your evidence
Prepare a one-page inventory of examples you can substantiate for large-scale development or consulting, hands-on platform engineering, and the listed tools. Note gaps rather than implying experience you do not have.
Rehearse a project walkthrough
Select a real project, if available, and prepare a five-minute account of your architecture decisions, code you contributed, deployment, outcomes, and how you led work across business, IT, infrastructure, and engineering teams.
Practice a cloud-native platform design
Sketch a hypothetical self-service AI platform with interoperable APIs, a Python/FastAPI service, container deployment, and an AWS or Azure path. Be ready to explain scaling, integration choices, and what you would take to an Enterprise Architecture review.
Prepare an agentic AI decision case
Outline how you would choose a model and framework, decide whether fine-tuning is warranted, evaluate an agentic workflow, and manage inference and model-lifecycle changes. Label this as a design exercise unless it reflects work you actually did.
Turn a process opportunity into an MVP plan
Draft a hypothetical current-state to future-state example: identify standard versus bespoke needs, define an MVP feasibility test, and specify what evidence would inform a broader automation decision.
Assemble delivery and quality artifacts
Bring redacted examples, if available, of a design artifact, Jira story, version-control or CI/CD approach, and quality checks. Prepare to explain risk ownership, sprint coordination, and collaboration with data scientists, ML engineers, software engineers, and QA.
Write focused questions for State Street
Ask how the Global AI Team selects self-service platform opportunities, what Enterprise Architecture expects in a design review, and how MVP feasibility and model quality are evaluated. These are questions to clarify the role, not assumed company practices.

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 of America. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…