Lead AI and Decision Engineer

What matters most

• Field decision-support AI for consequential use

  The role designs, develops, evaluates, and fields AI solutions that inform mission-critical decisions. Prepare to explain how a solution moved beyond a prototype and what evidence supported its use; deployment outside a lab is preferred, not required.

• Start with operational workflows

  MITRE calls for analyzing workflow and operational needs before building decision-support prototypes, including capabilities using large language models. Prepare an example of how users’ decisions shaped the prototype and the choice of AI techniques.

• Evaluate against sponsor requirements

  The role assesses AI solutions against sponsor requirements and produces actionable recommendations. Prepare to connect evaluation results, limitations, and a clear recommendation rather than presenting model performance alone.

• Lead interdisciplinary delivery

  The posting calls for technical leadership on cross-functional teams, project or small-team leadership, mentoring, and engagement with senior mission stakeholders. Prepare to describe how you set technical direction while incorporating operational and specialist input.

• Shape and communicate new R&D

  The engineer identifies and leads new AI-enhanced decision-systems research and conveys complex ideas through briefings or white papers. Prepare to explain a research opportunity, its relevance to a decision problem, and its implications for both expert and nontechnical audiences.

Likely challenges to think through

• Balancing useful decisions with uncertain AI outputs

  Likely preparation hypothesis: In high-stakes settings, a capable model may still produce outputs that are difficult to trust for a particular decision. Prepare to discuss how you would identify failure modes, evaluate against sponsor needs, and communicate limits on use.

• Choosing technology without losing the workflow

  Likely preparation hypothesis: An LLM-based approach may be appealing even when another technique or a simpler prototype better fits the operational need. Prepare to compare options using the decision workflow, user interaction, and evidence needed to justify the choice.

• Reconciling novel research with actionable delivery

  Likely preparation hypothesis: Proposing cutting-edge R&D while delivering useful sponsor recommendations creates a trade-off between exploration and near-term results. Prepare to explain how you would bound an experiment and turn its findings into a practical next step.

Topics to prepare

• Machine learning, deep learning and reinforcement learning

  Required hands-on areas. Revise how each learns, what decisions it can support, and its limitations; as an exercise, compare their suitability for one consequential decision problem.

• LLMs in decision support

  The role applies LLMs to sponsors’ decision-making challenges. As a revision exercise, sketch where an LLM could assist a workflow, where human judgment remains necessary, and how you would check the usefulness of its output.

• High-stakes decision-making gaps

  The posting asks for the ability to articulate challenges and identify gaps. As an exercise, take a hypothetical mission-critical decision and map its information gaps, operational constraints, and consequences of an incorrect recommendation.

• AI testing, evaluation and fielding

  The role evaluates solutions against sponsor requirements and fields AI systems; testing and evaluation experience is preferred. Revise the distinction between a promising prototype and a solution suitable for use outside a laboratory, then outline evidence needed for each.

• One additional technical specialty

  The qualifications require experience in at least one listed area beyond machine learning, such as decision science, human-machine teaming, automated planning, or knowledge-based reasoning. Choose an area you can substantiate and, as a revision exercise, explain how it would shape an AI-enhanced decision system.

Skills to revise

• Translate workflows into a prototype plan

  The role develops decision-support software prototypes from workflow and operational analysis. Practice turning a hypothetical sponsor workflow into a problem statement, proposed prototype capabilities, and a short list of assumptions to validate.

• Design an evaluation against requirements

  The posting calls for actionable recommendations based on sponsor requirements. Practice drafting a compact evaluation plan with requirements, measures, failure cases, and a recommendation justified by possible results.

• Frame an AI research opportunity

  The role proposes and executes novel R&D. Practice writing a one-page suggestion that states a decision-system gap, research question, proposed approach, and what result would warrant further work.

• Lead an interdisciplinary technical discussion

  Technical leadership, mentoring, and engagement with senior mission stakeholders are explicit responsibilities. Rehearse a meeting outline that elicits operational needs, resolves a technical trade-off, and assigns follow-up work across disciplines.

• Brief expert and nontechnical audiences

  The role communicates through briefings or white papers. Practice explaining the same proposed AI decision aid twice: once with technical evaluation details and once in plain language focused on the decision, limitations, and recommendation.

• Build a qualifications evidence sheet

  Map the posted experience-and-degree pathways, hands-on machine learning, deep learning, reinforcement learning, and at least one listed specialty to specific work you can substantiate. Mark any gaps rather than inventing evidence.

• Prepare a decision-support project walkthrough

  If you have a relevant project, outline the operational need, your technical decisions, prototype or deployed system, evaluation, stakeholder use, and lessons learned. The posting particularly values systems used outside a lab.

• Practice a high-stakes system design exercise

  For a hypothetical sponsor workflow, sketch how you would identify decision gaps, define requirements, prototype AI-assisted support, and keep human decision-makers involved. Treat this as interview practice, not a stated interview task.

• Draft an evaluation plan

  Prepare a one-page example showing how you would test an AI solution against sponsor requirements, examine consequential failure modes, and turn findings into actionable recommendations.

• Outline a research proposal

  Write a short, clearly hypothetical proposal for an AI-enhanced decision-system opportunity: the gap, research question, prototype approach, evaluation criteria, and potential operational value.

• Rehearse a leadership briefing

  Prepare a concise account of how you led an interdisciplinary effort, engaged senior stakeholders, or mentored technical staff, if applicable. Practice explaining the technical tradeoffs to both expert and nontechnical audiences.

• Review the stated work conditions

  Note that the posting lists McLean and Bedford, requires at least three on-site days weekly, and requires U.S. citizenship for clearance eligibility. No clearance is required to apply; the hired candidate must have or obtain DoD Top Secret/SCI within one year.

• Bring targeted employer questions

  Ask which sponsor decisions the team supports, how operational requirements and evaluation criteria are set, what fielding entails, and how a lead balances research, delivery, and mentoring.