AI for Engineering Professionals Capstone Exercise — AI for Engineering Professionals

What you'll learn

Apply skills from across this course in a single realistic multi-discipline project scenario
Produce a set of concrete, project-relevant deliverables using AI tools within a defined verification discipline
Mark exactly where independent, non-AI verification and licensed-engineer judgment take over from an AI draft

Engineers who bring AI into their practice well are not the ones who produce the fastest first draft. They are the ones who can produce a fast first draft and then draw an unambiguous, defensible line around exactly what still needs independent verification before it can leave the office. This capstone puts you in that position across four different categories of engineering work from this course.

You are a project engineer at a mid-size mechanical consulting firm, leading the mechanical scope on a warehouse-to-cold-storage conversion project. The client needs a preliminary technical package within five business days to support a financing decision. You will use AI to accelerate drafting across four deliverables while applying the verification discipline from this course to each one.

Capstone Exercise

Produce a Preliminary Technical Package for a Cold Storage Conversion

Context

Your firm is converting an existing 40,000-square-foot dry warehouse to a temperature-controlled cold storage facility (target: -10°F freezer zone, 35°F cooler zone). The client needs a preliminary technical package to support a financing decision in five business days. You have confirmed the existing structural and electrical capacity with the structural and electrical leads on your team; your scope is the mechanical (refrigeration and building envelope) package. This package is explicitly preliminary and will be superseded by full stamped design documents in a later phase, but the client will use it to make a real financing decision.

Your Task

Use an AI tool of your choice to produce first drafts of four deliverables: (1) A design basis memo summarizing the proposed refrigeration system approach, expected equipment categories, and the key assumptions driving the design (given: 40,000 sf, two temperature zones, target temperatures above); (2) A code and standards orientation list identifying which code families and standards likely govern this project (building code, mechanical code, refrigeration-specific standards such as ASHRAE 15/IIAR guidance, insulation/vapor barrier requirements) -- as an orientation list only, not verified citations; (3) A preliminary equipment specification outline for the primary refrigeration equipment, including the sections a full specification would need; (4) A one-page plain-language summary of the technical package for the client's financing team, who have no engineering background. For each of the four deliverables, write one annotation identifying: what in the AI draft is a placeholder or unverified claim that must be independently confirmed before this package is finalized, and what verification step (code lookup, calculation, standards register check, or licensed-engineer review) that specifically requires.

Your notes (optional)

Deliverable

A four-part preliminary technical package (design basis memo, code/standards orientation list, equipment specification outline, client summary) totaling 500 to 700 words, with one verification annotation per deliverable identifying the specific unverified claim and the verification step required before the package can be finalized into a stamped design phase.

Quick check

The capstone asks you to prompt for code families and topic areas rather than section numbers when building the standards orientation list. What should you do if the model supplies section numbers anyway?

Select one answer.

Key takeaways

AI can compress first-draft time significantly across design basis writing, code orientation, specification outlining, and client communication -- but a preliminary technical package used for a real financing decision still requires the same verification discipline as a stamped deliverable, applied at the appropriate stage.
Code and standards orientation lists are a starting point for research, not a compliance determination -- the annotation discipline from this capstone is the same discipline required before any AI-assisted code citation enters a real project document.
Plain-language client communication built from a technical design basis must preserve every assumption and open item -- a financing decision made on an over-simplified summary is exactly the scenario Lesson 6 warns against.
The annotation habit practiced in this capstone -- marking exactly what remains unverified and what verification step closes that gap -- is the professional habit this entire course has been building toward, and it does not stop being necessary once you leave this course.