Durable Engineering Skills

Summary

Peter Martin and Brian Douglas discuss eight engineering skills they believe will remain valuable as tools—including AI—continue to change. Their central point is that durable engineering ability is not just knowing how to use a particular technique or tool. It is knowing how to apply technical skills thoughtfully: identifying the real problem, judging whether an answer makes sense, understanding how systems fit together, and working effectively with other people.

They also discuss how education can help students develop these abilities through projects, realistic constraints, open-ended problem solving, and assessments that reveal students’ reasoning—not just their final answers.

Eight Durable Engineering Skills

1. Frame the problem before choosing a solution

• Start by clarifying what needs to be accomplished rather than reaching immediately for a familiar method.
• Examine available information, identify patterns, and ask what is known and unknown.
• Consider what outcome is needed and how success could be judged.
• In teaching, instructors can present a problem or data first, invite students to reason about it, and then introduce tools such as linear regression as possible ways to solve it.
• This helps engineers avoid becoming stuck when they do not recognize a ready-made solution.

2. Scope solutions around real needs and constraints

• A technically impressive solution is not necessarily the right one. It must address the customer’s actual need.
• Consider practical limits such as time, budget, performance requirements, materials, power, weight, and production scale.
• Aim for the level of performance that is useful—not automatically the maximum possible performance.
• Ask how a design would change if it had to be built quickly, cheaply, or thousands of times.
• Education can make these trade-offs visible by adding constraints to projects or asking students how a one-off prototype would need to change for production.

3. Use physical intuition to sanity-check results

• Before accepting a simulation or calculation, ask whether the result makes sense for the physical system.
• Develop an expectation about the system’s behavior, then compare the computed answer with that expectation.
• For example, an unnecessarily high-bandwidth controller might react to noise rather than meaningful disturbances. A simulation result that conflicts with basic physical expectations may indicate a modeling or implementation error.
• Students should be encouraged to anticipate likely outcomes before calculating them and to explain why their design choices—such as bandwidth or sampling rate—are reasonable.

4. Think in systems

• Understand how a larger system is organized, including its components, interfaces, and layers.
• A new engineer may be asked to modify existing code, add a sensor, or change a system’s behavior—not build an entire system from scratch.
• Being able to identify how sensing, low-level control, higher-level decisions, software, and physical components relate helps engineers make changes without losing sight of the whole.
• Project work with realistic, multi-part systems can help students develop this perspective.

5. Exercise judgment about models

• Models are essential for simulation, prediction, design, verification, and control, but no model captures every detail.
• Decide what to include or omit based on the purpose of the model and the question being answered.
• A mathematically sound model may describe an idealized system rather than real hardware, which can include friction, backlash, sensor filtering, quantization, or other effects.
• A useful approach is to begin with a model that is adequate for the task, then tune or refine it as real-world behavior reveals what matters.
• Students should learn not only how to build models, but also why a model is being built and what assumptions it makes.

6. Build engineering literacy across disciplines

• Most engineering work is collaborative and multidisciplinary. Engineers benefit from understanding the concerns and terminology of colleagues in other fields.
• For example, mechanical, electrical, software, controls, applications, and production specialists need to communicate requirements and understand how decisions in one area affect another.
• Engineers should be able to explain their own needs while also anticipating and advocating for other team members’ concerns.
• This matters in large organizations with specialized groups, as well as in small companies or startups where individuals may need to cover several disciplines.
• Engineering education can introduce students to what other disciplines do and the problems they address, even when students are not training to become specialists in those fields.

7. Follow sound engineering workflows

Do not move directly from an idea or prompt to implementation. Use a deliberate process:

1. Define the problem and what the solution is supposed to accomplish.
2. Establish requirements or specifications where appropriate.
3. Break a complex system into manageable parts.
4. Choose models, tools, code, and test methods suited to the task.
5. Verify components and validate whether the finished system meets its intended goals.
6. Troubleshoot, revise, and document the work.
7. Consider maintenance, updates, and how the solution will behave as its environment changes.

AI and other tools can speed up execution, but they do not replace the judgment needed to ask the right questions or decide whether a proposed solution is appropriate. A generated answer or code sample is not, by itself, evidence that the system is correct, safe, maintainable, or fit for purpose.

8. Make mistakes, learn from them, and communicate honestly

• Mistakes are part of engineering and can be useful when they reveal weaknesses before a customer encounters them.
• Engineers should document what happened, identify how to prevent recurrence, and limit the scope and consequences of failures.
• New engineers should feel able to raise concerns, contribute ideas, and take responsibility for their area of expertise.
• It is better to clearly say “I don’t know” than to pretend to understand something; openness makes it possible to learn and draw on colleagues’ expertise.
• Confidence does not mean knowing everything. It means being willing to take on appropriate challenges, learn from outcomes, and ask for help when needed.

Educational Implications

The speakers favor approaches that make students practice engineering judgment, not just reproduce a taught procedure. They suggest:

• Using project-based learning to give students open-ended problems and opportunities to experiment.
• Introducing realistic requirements and constraints, such as cost, schedule, performance limits, and production scale.
• Asking students to explain their reasoning, assumptions, and design decisions.
• Using discussions and oral presentations to assess whether students understand their work, including when they use computational tools or AI.
• Teaching students to connect technical methods to practical goals and to recognize that learning and iteration are ongoing parts of engineering.

Speakers and Sources Featured

• Peter Martin — Quanser; discusses engineering experience, robotics, mechatronics, and industrial practice.
• Brian Douglas — MathWorks; discusses controls, engineering education, and technical communication.

Organizations and institutions mentioned in examples: Quanser, MathWorks, the University of Toronto, and Aalborg University in Denmark.