New opportunity

AI/Machine Learning Engineer

Fordham University · New York, United States

About this role

Reporting to the Assistant Vice President of Enterprise AI, the AI/Machine Learning Engineer leads technical implementation of practical AI and machine learning solutions across Fordham University. The role combines AI agents supporting the student journey with predictive models for institutional planning and student outcomes. It works with university leadership, IT, data governance, academic and administrative teams, and external vendors. The position is hybrid across the Rose Hill and Lincoln Center campuses, with 35 scheduled hours per week.

Responsibilities

Design and deploy multi-agent systems for reasoning, tool use and autonomous problem-solving across recruitment, learning and development, career services, and alumni relations. Architect agents using retrieval-augmented generation (RAG), Model Context Protocol and operational data.
Develop and maintain production-oriented predictive models for student success, enrollment planning, operational efficiency and early identification of students who may benefit from support. Build workflow automation, decision-support tools and AI-assisted administrative processes.
Build RAG workflows, vector database integrations and other AI application components using approved platforms and institutional data. Develop MLOps frameworks to monitor model drift and performance in compliance with AI governance policy.
Assess feasibility, document solution architecture and recommend scalable implementation approaches, with attention to ethics, privacy, security, FERPA, GDPR and applicable governance standards.

Required qualifications

A master’s degree in Computer Science, AI/Data Science, Engineering or a related field; a minimum of 5 years of experience in machine learning or data science, including deployment of predictive models in production; deep proficiency with LLMs and development and deployment of autonomous AI agents, for example using LangChain and CrewAI; strong statistical modeling and classification skills; experience building RAG pipelines and working with vector databases; and strong Python coding and MLOps skills. Candidates should understand and support the goals of Jesuit education.

Preferred qualification: A Ph.D. in Computer Science, Engineering or a related field. The starting salary range is $128,000–$160,000, commensurate with qualifications, experience and skills. Fordham offers medical, dental and vision insurance, retirement plans, tuition remission and time off, among other benefits.

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

Skills for this role

AI agentsMulti-agent systemsMachine learningPredictive modelingLarge Language Models (LLMs)Retrieval-Augmented Generation (RAG)Vector databasesModel Context ProtocolLangChainCrewAIStatistical modelingClassification algorithmsPythonMLOpsModel drift monitoringAI governanceFERPAGDPR

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

Deploy agents across the student journey
The posting calls for multi-agent systems with reasoning and tool use across recruitment, learning and development, career services, and alumni relations. Prepare to explain an agent architecture, including how it would use RAG, Model Context Protocol, and operational data.
Deliver actionable predictive models
The role develops and maintains models for student success, enrollment planning, operational efficiency, and identifying students who may benefit from timely support. Prepare to discuss how a production model’s outputs would inform a decision or intervention.
Build maintainable AI applications
The posting specifies RAG workflows, vector database integrations, approved platforms, and MLOps monitoring of drift and performance. Prepare a concrete account of deployment, evaluation, and ongoing maintenance rather than a prototype alone.
Make governance part of the architecture
The engineer must translate institutional needs into secure, ethical, privacy-compliant solutions, with attention to FERPA, GDPR, and AI governance policy. Prepare to describe how data access and privacy constraints shape technical choices.
Turn institutional needs into feasible plans
As technical implementation lead, the role assesses feasibility, documents architecture, supports the AVP’s implementation priorities, and liaises with leadership, technical teams, and vendors. Prepare to explain how you would compare approaches and communicate trade-offs to different stakeholders.

Likely challenges to think through

Agent autonomy versus controlled access
Preparation hypothesis: Agents that use tools and institutional data may be harder to keep within appropriate access and action boundaries. Prepare to explain where you would constrain tool use, require human review, and test behavior before deployment; the posting does not identify a current failure.
Early support signals versus reliable interpretation
Preparation hypothesis: A model intended to identify students who may benefit from support must be useful without treating predictions as certainties. Prepare to discuss validation, drift monitoring, and how decision-makers should interpret outputs.
Campus-wide scale versus local needs
Preparation hypothesis: Recruitment, academic, administrative, and alumni use cases may differ even when they share an enterprise AI roadmap. Prepare to show how you would assess feasibility and prioritize reusable components while preserving each use case’s requirements.

Practice your answers

Technical practice

How would you architect a multi-agent system to support a stage of the student journey?

What to coverSuggested answer focus: explain how agents would divide reasoning and tasks, use tools, and hand work off while keeping a clear path for human review. Discuss reliability and scalability for a use case such as recruitment or career services.

How would you connect an autonomous agent to operational data and tools?

What to coverSuggested answer focus: describe where Model Context Protocol could fit and how you would limit tool permissions, validate actions, and record what an agent did. Distinguish useful automation from actions that should require approval.

How would you build a RAG workflow using institutional data?

What to coverSuggested answer focus: cover document preparation, retrieval, vector database integration, and how responses would remain grounded in retrieved material. Explain how access controls and data updates would affect the design.

How would you evaluate a RAG-based agent before deployment?

What to coverSuggested answer focus: propose tests for retrieval relevance, answer accuracy, tool-use behavior, and failures on representative tasks. Explain the trade-off between broader autonomy and predictable, reviewable results.

How would you develop a model to identify students who may benefit from timely support?

What to coverSuggested answer focus: discuss outcome definition, statistical or classification approaches, data leakage, and evaluation against the intended support decision. Address false positives, missed students, and the need to assess disparate impacts rather than treating a score as a decision.

How would you approach a predictive model for enrollment planning?

What to coverSuggested answer focus: explain how you would establish a baseline, choose appropriate inputs and validation periods, and communicate uncertainty to planners. Discuss when a simpler interpretable model might be preferable to a more complex one.

How would you maintain a production predictive model in Python?

What to coverSuggested answer focus: describe reproducible training and deployment, performance monitoring, and how you would detect and investigate drift. Explain what evidence would prompt retraining, rollback, or a review of the underlying decision process.

How would you design an AI application around privacy and governance requirements?

What to coverSuggested answer focus: explain how you would assess institutional data use, security, and applicable FERPA, GDPR, and AI governance obligations before implementation. Discuss data minimization, access boundaries, and what you would document for review.

When would AI-assisted workflow automation be preferable to a decision-support tool?

What to coverSuggested answer focus: compare the operational task, error consequences, and need for human judgment in an academic or administrative process. Explain how you would measure whether either approach improves operations without obscuring accountability.

How would you assess the feasibility of a proposed enterprise AI solution?

What to coverSuggested answer focus: walk through data availability, approved platforms, architecture, integration effort, governance constraints, and scalability. Explain how you would document options and trade-offs to support implementation priorities set by the Assistant Vice President of Enterprise AI.

Behavioural practice

Tell me about a time you aligned technical and nontechnical stakeholders on an AI or data solution.

What to coverSuggested answer focus: use a real situation involving differing needs, explain how you translated them into a workable approach, and state the result. The role calls for liaison work across leadership, technical teams, academic or administrative stakeholders, and vendors.

Tell me about a time you challenged or changed a proposed use of sensitive data.

What to coverSuggested answer focus: describe the situation, the privacy or ethical concern you identified, the action you took with relevant stakeholders, and the result. Keep the example grounded in what you actually did rather than assuming a particular university policy.

Tell me about a time you had to prioritize competing implementation requests.

What to coverSuggested answer focus: explain the competing needs, how you assessed feasibility and impact, what you recommended, and the outcome. This is relevant to the role’s contribution to a technical roadmap and implementation priorities.

Tell me about a production model or AI application that did not perform as expected.

What to coverSuggested answer focus: give the situation, how you detected and investigated the problem, the action you took, and the result. Include how you communicated uncertainty or operational impact to the people relying on it.

Tell me about a time you balanced a technology objective with an organization’s mission.

What to coverSuggested answer focus: use a real situation, describe the choice you made and why, and explain the result. For this role, be prepared to discuss how your approach to AI would support the stated goals of Jesuit education without claiming experience you do not have.

Sharpen your knowledge

Topics to prepare

LLM and multi-agent architecture
Map how agents could divide reasoning, tool use and handoffs across a student-journey task; identify where human review would be prudent. The posting explicitly calls for deployed autonomous multi-agent systems; human-review checkpoints are a preparation suggestion, not a stated requirement.
RAG, vector retrieval and Model Context Protocol
Sketch how an agent would retrieve relevant institutional content, use a vector database and access operational data through Model Context Protocol. Explain what each component contributes. The posting names all three in its agent architecture.
Statistical modeling and classification
Compare two classification approaches for identifying students who may benefit from support, including how you would assess errors and whether results are actionable. The role requires statistical modeling and classification for student outcomes and planning.
Model performance and drift
Work through a hypothetical change in enrollment patterns: distinguish a change in input data from a drop in predictive performance and propose what to monitor. The posting requires MLOps frameworks that monitor drift and performance.
Governance and educational context
Review a hypothetical student-data use case for privacy, security, ethics, FERPA/GDPR considerations and the goals of Jesuit education. These are stated considerations for the role; the hypothetical review is a preparation exercise, not a description of Fordham policy.

Skills to revise

Implement a predictive model in Python
Using synthetic student-outcome data, code a classification workflow and produce an evaluation summary that explains its limits for decision support. Python coding and production-deployed predictive models are explicit requirements.
Build a small RAG workflow
With non-sensitive sample documents, implement retrieval through a vector database and check whether generated answers reflect the retrieved material. The posting calls for RAG pipelines and vector database integrations using approved platforms and institutional data.
Prototype agent coordination
Create a small demonstration in which agents hand off a student-services request and use a permitted tool; test a failed handoff. LangChain or CrewAI may be used as examples named in the posting, not as mandatory frameworks.
Exercise a monitoring workflow
Add checks to a sample predictive-model workflow that record performance over time and flag a simulated drift scenario. This directly rehearses the stated MLOps and model-monitoring responsibilities.
Document an implementation approach
Write a one-page architecture and feasibility brief for a proposed student-journey agent, covering scalability, privacy and stakeholder questions. The role assesses feasibility, documents architecture and works with university leadership, IT, data governance and academic and administrative teams.

Your preparation checklist

Map your evidence to the stated qualifications
Prepare a one-page reference for your résumé and interview covering the required master’s degree, at least five years in ML or data science, production predictive-model deployments, Python, LLM agents, RAG, vector databases, and MLOps. Distinguish direct experience from areas you would need to learn.
Rehearse an agent-system walkthrough
Select a relevant project and explain its architecture, tool use, deployment, and safeguards. Connect it to the role’s stated interest in multi-agent systems for recruitment, learning and development, career services, or alumni relations; be ready to discuss RAG and Model Context Protocol where applicable.
Document a production modeling example
Prepare a concise walkthrough of a predictive model: the decision it supported, statistical or classification approach, production deployment, performance measurement, drift monitoring, and maintenance. Relate the discussion to student success or enrollment planning without claiming experience with university data unless you have it.
Practice a hypothetical university AI design exercise
Sketch an agent or decision-support workflow using approved platforms and institutional data. Show the architecture, feasibility trade-offs, human review points, and how you would address privacy, security, ethics, FERPA/GDPR, and AI governance; present it as a proposed design, not a Fordham project.
Prepare a cross-functional delivery story
Choose an example of translating stakeholder needs into a technical implementation. Explain how you aligned technical teams, decision-makers, and any vendors, reflecting the role’s liaison work with university leadership, IT, data governance, and academic and administrative teams.
Tailor the required application materials
Prepare the résumé and cover letter listed in the posting. Use specific, supportable examples of technical implementation leadership and explain your understanding of and commitment to the goals of Jesuit education.
Write questions about priorities and constraints
Ask how the AVP prioritizes agent and predictive-model work, which institutional data and approved platforms are available, how AI governance reviews operate, and what outcomes would define success for the technical implementation lead.

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…