Stop Building Things Nobody Asked For
Here is a pattern that plays out in engineering programs all over the world. A team of brilliant students picks a project, spends months building an impressive technical solution, presents it at the end of the semester, and nobody ever uses it. The tech was solid. The execution was clean. But the problem it solved? Either it did not exist, or it existed differently than the team assumed.
Design thinking flips that entire script. Instead of starting with “what can we build?”, you start with “who is struggling, and what does their struggle actually look like?” You go talk to real people. You watch how they work, eat, commute, study, or recover. You find the messy, unglamorous friction points that textbooks never mention. And then you build something that actually helps. The same principle applies when exploring AI-based mini project ideas for engineering students: the strongest projects are not just technically interesting; they connect tools like machine learning, computer vision, NLP, or IoT to a problem people genuinely face. That is where a student project starts becoming something more than an academic exercise.
This article gives you 25 project ideas organized around real-world problems that affect real people. Each one is grounded in a scenario you could walk into tomorrow and start researching. None of them require a PhD or a $50,000 budget. All of them will stretch your thinking in ways that pure technical coursework never does.
| Why this matters for your career: Design thinking is not a nice-to-have anymore.A 2024 meta-analysis of 25 studies found that design thinking positively affects student learning with a significant effect size (r = 0.436, p < 0.001). IEEE identifies the design thinking plus agile sprint combination as the dominant model in modern product development. And 86% of employers now rank creative problem-solving as a top hiring priority. Engineering grads who can show empathy-driven project work outperform those who cannot. |
Why design thinking skills matter: employer priorities in 2026.
A 60-Second Refresher on How Design Thinking Works
If you have seen this before, skip ahead. If you have not, here is the entire framework in five phases. It is not a straight line. You will bounce between phases constantly, and that is the point.
The five phases of design thinking. You will revisit each one multiple times per project.
Empathize: Go talk to the people who actually have the problem. Watch them. Ask open-ended questions. Do not pitch solutions yet. Just listen.
Define: Take everything you heard and boil it down to one clear problem statement. “How might we [do X] for [these people] so that [this outcome happens]?”
Ideate: Generate as many possible solutions as you can. Quantity over quality at this stage. No idea is too dumb. Some of the dumbest ideas lead to the best solutions.
Prototype: Build the simplest possible version of your best idea. Cardboard, paper, Figma mockups, 3D prints, a landing page. It just needs to be real enough that someone can react to it.
Test: Put your prototype in front of real users. Watch what they do (not what they say they would do). Take notes. Go back to any earlier phase as needed.
How the 25 project ideas distribute across five real-world impact areas.
Each domain challenges different design thinking muscles. Healthcare demands deep empathy. Urban projects need more tech integration.
Healthcare and Accessibility
These projects deal with people at their most vulnerable. That makes the empathy phase harder, but it also makes the impact more meaningful. You will learn quickly that solutions designed without talking to patients, caregivers, or people with disabilities almost always miss the mark.
1. Medicine Reminder for Elderly People Living Alone
4-6 weeks | Empathy-heavy
The scenario: Your grandmother takes five different medications at different times of day. She sometimes forgets, sometimes double-doses, and the pill organizer her doctor recommended is confusing. Her eyesight is not great, and she does not like phone apps.
What you actually build: A physical or hybrid reminder system designed through interviews with 5-10 elderly people about their actual medication routines, barriers, and preferences. The prototype could be a redesigned pill organizer, a simple alert device, or a low-tech visual system. The form follows the research, not the other way around.
Why this is real: Medication non-adherence costs healthcare systems billions and causes preventable hospitalizations. Existing solutions are overwhelmingly designed by young people who never tested them with actual elderly users.
Skills practiced: User interviews with sensitive populations, accessibility-first design, physical prototyping, iterative testing with non-tech-savvy users.
2. Wheelchair-Accessible Campus Navigation
6-8 weeks | High field research
The scenario: A student using a wheelchair discovers that the “accessible route” shown on the campus map includes a ramp that has been broken for three months, a door that requires badge access she does not have, and a detour that adds 12 minutes to a 3-minute walk.
What you actually build: An audit-based navigation tool that maps actual accessible routes on your campus (not the official ones, the real ones). Interview wheelchair users and mobility-impaired students. Walk every route yourself. Build a map or app prototype that shows what actually works today.
Why this is real: Every university claims ADA compliance. Very few have tested their routes with actual wheelchair users in the last year. The gap between official accessibility and experienced accessibility is enormous.
Skills practiced: Accessibility auditing, participatory design, spatial mapping, stakeholder interviews (facilities staff, students, administration).
3. Mental Health Check-In Tool for Engineering Students
6-8 weeks | Ethics-sensitive
The scenario: It is week 10 of the semester. Three exams in four days. A classmate who was always in the front row stopped showing up. Nobody checked in because everyone assumes someone else will.
What you actually build: A peer-driven check-in system designed with input from your campus counseling center and students who have experienced burnout. This could be a physical system (a simple “how are you doing?” board in the common area), a digital tool, or a structured peer support process. Research from Frontiers in Psychiatry (2025) found that AI chatbots reduced student depression scores by 22%, so there is real evidence behind technology-assisted mental health support.
Why this is real: Engineering programs have some of the highest burnout rates in higher education. Most mental health tools are designed generically. Very few are designed specifically for the unique pressures of engineering coursework.
Skills practiced: Sensitive user research (ethics matter here), co-design with counselors, privacy-first design, balancing technology with human connection.
4. Low-Cost Prosthetic Hand for Developing Regions
8-12 weeks | High technical + high empathy
The scenario: A 14-year-old in a rural area lost three fingers in a farming accident. The nearest prosthetics clinic is 200 kilometers away. A conventional prosthetic costs $15,000. The family earns $3,000 per year.
What you actually build: Research existing open-source prosthetic designs (like e-NABLE), interview local healthcare workers or NGO staff about real constraints (materials available, repair capabilities, cultural attitudes toward prosthetics), and adapt or redesign a 3D-printable hand that costs $300-500 and can be maintained locally.
Why this is real: 3D-printed prosthetics at $300-500 versus $15,000 for conventional options are already changing lives. The design challenge is not the printing. It is understanding the user context: what activities matter, what materials are locally available, and what maintenance is realistic.
Skills practiced: Contextual design for resource-constrained environments, open-source collaboration, 3D modeling/printing, cost optimization.
5. Redesigning Hospital Discharge Instructions
4-6 weeks | Information design focus
The scenario: A patient is sent home after minor surgery with a packet of instructions printed in 10-point font, full of medical jargon, organized in no particular order. They leave confused, call the clinic twice, and end up back in the ER because they did not understand a wound care instruction.
What you actually build: Interview recently discharged patients and ER staff about the most common confusion points. Redesign the discharge packet using plain language, visual instructions, clear hierarchy, and a simple checklist format. Test it with real patients.
Why this is real: Nearly 20% of patients are readmitted within 30 days, and a significant portion of those readmissions trace back to poor understanding of discharge instructions. This is a solvable design problem.
Skills practiced: Health literacy research, information design, plain-language writing, patient-centered testing, visual communication.
6. Sensory-Friendly Waiting Room Design
5-7 weeks | Multi-stakeholder
The scenario: A child with autism spectrum disorder is brought to a dental clinic. The waiting room has fluorescent lights, a TV blaring news, plastic chairs, and a pile of old magazines. The child has a meltdown before the appointment even starts.
What you actually build: Research sensory processing challenges through interviews with parents, occupational therapists, and (where appropriate) neurodiverse individuals. Design a waiting room environment or a portable “sensory kit” that clinics could adopt. Prototype the kit or create a detailed design specification with material samples.
Why this is real: Sensory-friendly design is increasingly recognized in healthcare architecture, but very few clinics have implemented it. The gap between awareness and action is your design opportunity.
Skills practiced: Inclusive design research, environmental design, material selection, stakeholder mapping (patients, parents, clinic owners, therapists).
Sustainability and Environment
Sustainability projects are everywhere in engineering courses. Most of them are terrible. They start with a technology (“let us build a solar-powered thing”) instead of a human problem (“why do people in this neighborhood throw recyclables in the trash?”). The projects below force you to start with the people.
7. Reducing Food Waste in Your Campus Cafeteria
6-8 weeks | Measurable impact
The scenario: The cafeteria throws away 200 kg of food every week. Some of it is overcooked, some of it is food nobody ordered, and a lot of it is perfectly good food that students took but did not finish.
What you actually build: Spend a week observing cafeteria operations. Interview kitchen staff, dining hall managers, and students. Map where the waste actually happens (preparation, serving line, or trays). Design and test an intervention: portion size options, a pre-order system, a “rescue shelf” for untouched food, or a feedback mechanism that connects students to the kitchen.
Why this is real: Campus food waste is measurable, local, and fixable. You can track your impact in kilograms within weeks, which makes this one of the most satisfying design thinking projects you can do.
Skills practiced: Observational research, process mapping, behavioral design, measurement and iteration, stakeholder alignment.
8. Smart Irrigation for Small Farmers
8-10 weeks | Cross-domain
The scenario: A smallholder farmer in a semi-arid region waters her crops by flood irrigation because that is what her parents did. She uses three times more water than her crops need, her yields are inconsistent, and drought years devastate her income.
What you actually build: Research existing smart irrigation systems, interview agricultural extension workers or small farmers (even via video call), and design a low-cost, low-tech irrigation controller that saves 35-45% water. The prototype could use soil moisture sensors and a simple relay, or it could be entirely manual with a redesigned scheduling guide.
Why this is real: Smart irrigation systems saving 35-45% water are documented in peer-reviewed research (Springer Nature). The challenge is making them affordable and usable for farmers who have never used a sensor.
Skills practiced: Contextual design, IoT basics (if electronic), cost-sensitive prototyping, agricultural domain knowledge.
9. Neighborhood Recycling Behavior Change
4-6 weeks | Behavior design
The scenario: Your apartment complex has recycling bins, but 60% of the contents are contaminated with non-recyclable items. The recycling gets sent to the landfill anyway.
What you actually build: Interview residents about their recycling habits and confusion points. Observe the bin area. Redesign the bin signage, placement, or the entire system. Test whether your changes reduce contamination rates over 2-3 weeks.
Why this is real: Recycling contamination is a behavioral design problem, not a technology problem. Most contamination happens because people genuinely do not know what goes where. Signage is almost always poorly designed.
Skills practiced: Behavioral research, signage and information design, A/B testing in the real world, measuring behavior change.
10. E-Waste Collection System for College Campuses
4-6 weeks | Service design focus
The scenario: Students upgrade phones and laptops every 2-3 years. The old devices go in a drawer, get thrown in the trash, or sit in a box in the closet. The campus has no convenient e-waste collection point.
What you actually build: Survey students about their electronic disposal habits. Design a collection system that fits into the rhythms of campus life (maybe tied to move-out week, or placed in dorm lobbies). Include a clear communication strategy about what happens to the collected devices.
Why this is real: E-waste is one of the fastest-growing waste streams globally. Most campus e-waste programs fail not because of logistics but because of awareness and convenience. This is a pure design problem.
Skills practiced: Survey design, service design (not just product design), logistics planning, communication design.
11. Energy Audit Toolkit for Student Housing
4-5 weeks | Toolkit design
The scenario: A student’s electricity bill tripled in winter. They have no idea why. Is it the ancient space heater? The window that does not close properly? The computer that never gets turned off?
What you actually build: Design a simple, shareable energy audit toolkit that any student can use to find and fix energy waste in their housing. Test it with 5-10 students in different types of housing. Iterate based on what they actually found useful versus what they ignored.
Why this is real: Energy audits for commercial buildings are well-established. Energy audits for student apartments (shared, poorly maintained, with no control over major systems) are a completely different problem.
Skills practiced: Information design, worksheet/tool design, energy literacy research, user testing in diverse living situations.
Education and Learning
You are engineering students designing for other students. That gives you a massive advantage: you already know the users. But it also creates a trap. You think you understand the problem because you live it. Design thinking forces you to check that assumption by talking to people whose experience might be very different from yours.
12. Peer Tutoring Matching System
5-7 weeks | Service + social design
The scenario: A first-year student is struggling with thermodynamics. They do not want to go to office hours (too intimidating), and the tutoring center has a two-week wait. A third-year student two doors down aced that exact class and would happily help for free.
What you actually build: Interview students on both sides (those who need help and those who could provide it). Map the barriers to connection. Design a matching system that is low-friction, builds trust, and does not feel transactional. Test a prototype with a small pilot group.
Why this is real: Peer learning is one of the most effective educational interventions, but most universities leave it to chance. The design challenge is social, not technical.
Skills practiced: Two-sided user research, trust design, service blueprinting, pilot testing.
13. Lab Equipment Sharing Platform
5-7 weeks | Organizational design
The scenario: Three different research groups in the engineering building own the same $40,000 oscilloscope. Two of them sit idle 80% of the time. A fourth group needs one desperately but has no budget.
What you actually build: Interview lab managers and researchers about equipment usage patterns, sharing barriers, and maintenance concerns. Design a booking and sharing system that addresses the real objections (damage, availability, calibration). Prototype the workflow.
Why this is real: Equipment underutilization in universities is a well-documented problem. Sharing platforms exist for cars, homes, and office space. Why not lab equipment?
Skills practiced: Stakeholder interviews with competing interests, workflow design, trust and accountability systems.
14. First-Generation Student Onboarding Guide
4-6 weeks | Equity-focused
The scenario: A first-generation college student shows up to engineering orientation with no context. They do not know what “office hours” means, how to read a syllabus, what a TA does, or that “study groups” are a thing. Nobody explains these unwritten rules because everyone assumes you already know.
What you actually build: Interview first-generation engineering students (sophomores and above, so they can reflect on what they wish they had known). Design an onboarding resource that addresses the real gaps, not the official orientation content.
Why this is real: First-generation students drop out of engineering at significantly higher rates than their peers. The biggest factors are not academic. They are navigational and social.
Skills practiced: Empathic research with underrepresented populations, content design, testing for cultural relevance and tone.
15. Study Space Finder for Exam Season
4-6 weeks | High testability
The scenario: Every seat in the library is taken. The coffee shop is too loud. The empty classroom you used last week is locked. It is 10 PM, you have an exam at 8 AM, and you are carrying your laptop and three textbooks looking for a flat surface.
What you actually build: Map every possible study space on campus (including unofficial ones). Survey students about what they need in a study space (quiet, outlets, group-friendly, food-allowed). Design a real-time or crowdsourced availability system and test it during exam week.
Why this is real: Study space is a genuine pain point at every university. The problem is not a shortage of space. It is a shortage of information about available space.
Skills practiced: Environmental scanning, crowdsourcing design, information architecture, real-world stress testing.
16. Workshop Safety Training Redesign
5-7 weeks | Constrained design
The scenario: The mandatory workshop safety video is 45 minutes of someone reading slides. Nobody pays attention. Two students got minor injuries this semester because they did not remember the correct procedure for a specific machine.
What you actually build: Observe current safety training sessions. Interview students about what they actually remember (and what they ignored). Interview workshop technicians about the most common safety mistakes. Redesign the training using principles of active learning, visual communication, and just-in-time reference materials.
Why this is real: Safety training in engineering workshops is almost universally boring and ineffective. The design challenge is making critical information stick when people are not motivated to learn it.
Skills practiced: Learning design, visual instruction design, behavior change, regulatory constraints (you still have to cover everything mandated).
Urban Living and Mobility
Cities are design problems at scale. Every frustration you feel as a commuter, pedestrian, or apartment dweller is a problem someone can solve. These projects put you in the shoes of people navigating systems that were designed decades ago for different populations.
17. Last-Mile Delivery for Walkable Neighborhoods
6-8 weeks | Systems-level thinking
The scenario: Delivery trucks block bike lanes, double-park, and idle for 20 minutes while drivers carry packages up five flights of stairs. The neighborhood was designed before online shopping existed.
What you actually build: Observe delivery patterns in a specific neighborhood. Interview delivery drivers, residents, and local businesses. Design a last-mile solution: neighborhood locker systems, cargo bike hubs, scheduled delivery windows, or something nobody has tried yet.
Why this is real: Last-mile delivery is the most expensive and environmentally damaging part of the logistics chain. Urban areas designed for foot traffic are getting crushed by package volume.
Skills practiced: Observational research, multi-stakeholder mapping, urban systems thinking, feasibility analysis.
18. Safe Walking Routes for Children
4-6 weeks | Community-facing
The scenario: A parent drives their kid 800 meters to school because the direct walking route has no sidewalk for one block, a blind intersection, and a stretch with no lighting.
What you actually build: Walk every possible route to a local school during school hours. Interview parents, teachers, and crossing guards. Map the specific safety concerns (not general ones). Design improvements and present them to the school or neighborhood board.
Why this is real: School walk zones are determined by distance, not safety. A child who lives 500 meters from school on a dangerous road gets no bus service. The assumption is that “close” equals “safe.”
Skills practiced: Route auditing, safety assessment, community engagement, presenting design recommendations to decision-makers.
19. Public Restroom Finder for People with Medical Conditions
6-8 weeks | Dignity-centered design
The scenario: A person with Crohn’s disease avoids leaving home for more than an hour because they cannot reliably find a clean, accessible public restroom. Existing apps show locations but not cleanliness, accessibility, or whether a code is required.
What you actually build: Interview people with IBD, IBS, or other conditions that require urgent restroom access. Map real restroom data in a specific neighborhood. Design an app, a card system, or a partnership program that solves the actual barriers (not just location, but access, cleanliness, and dignity).
Why this is real: Restroom access is a genuine barrier to independence for millions of people. It is a problem nobody wants to talk about, which means almost nobody is designing for it.
Skills practiced: Sensitive user research, dignity-preserving design, data collection in public spaces, partnership design.
20. Noise Pollution Mapping and Intervention
6-8 weeks | Data + community
The scenario: A residential street next to a highway has noise levels that exceed WHO guidelines. Residents report sleep problems, stress, and difficulty concentrating. The highway is not going anywhere.
What you actually build: Measure actual noise levels at different times and locations using smartphone decibel meters. Interview affected residents about their experience and coping mechanisms. Design interventions (sound barriers, vegetation buffers, white noise systems, or community advocacy materials) that are realistic for the context.
Why this is real: Noise pollution is the “invisible pollutant.” It correlates with cardiovascular disease, cognitive impairment in children, and mental health issues. Engineers rarely design for it because it is not visible.
Skills practiced: Environmental measurement, data visualization, intervention design, community-engaged research.
21. Micro-Park Design for Underused Urban Spaces
5-7 weeks | Spatial + community
The scenario: There is a 12×20-meter strip of land between two buildings in a dense neighborhood. It is currently a parking lot for four cars. Five hundred people walk past it daily. None of them have anywhere to sit.
What you actually build: Survey the neighborhood about what they would use the space for. Research micro-park precedents globally. Design a space that serves the community’s actual needs (seating? shade? play space? phone charging? all of the above?) and present it as a proposal.
Why this is real: Micro-parks (pocket parks, parklets) are one of the most cost-effective urban interventions available. Most cities have unused spaces that could become community assets with thoughtful design.
Skills practiced: Community needs assessment, spatial design, precedent research, feasibility presentation.
Food and Water Security
Food and water are so fundamental that it is easy to forget they are design problems. Every step from production to consumption involves systems that were designed by someone, and many of those designs are failing the people who need them most.
22. Water Quality Testing Kit for Rural Communities
6-8 weeks | Communication design
The scenario: A family draws water from a well they have used for generations. They do not know if the water is safe. Testing at a lab costs money they do not have and takes a week to get results.
What you actually build: Research existing water testing methods and their cost, accuracy, and usability. Interview public health workers or community leaders about water concerns. Design a testing kit (or a testing process) that is affordable, gives results quickly, and communicates results in a way non-specialists can understand.
Why this is real: Unsafe drinking water affects billions of people globally. Low-cost test kits exist, but the design challenge is making the results actionable for people without scientific training.
Skills practiced: Scientific communication design, cost-constrained engineering, cross-cultural usability testing.
23. Community Fridge Network Design
6-8 weeks | Social infrastructure
The scenario: A neighborhood has both food waste (restaurants discarding unsold meals) and food insecurity (families skipping meals). The food bank is across town and only open on Tuesdays.
What you actually build: Research existing community fridge models (they exist in many cities). Interview potential donors (restaurants, bakeries, grocery stores), potential users, and health regulators. Design a community fridge system that addresses the logistical, regulatory, and social barriers.
Why this is real: Community fridges work. The design challenge is not the fridge itself. It is the trust network, food safety protocols, and community ownership model around it.
Skills practiced: Service design, regulatory research, community organizing, trust-building design.
24. Nutrition Label Redesign for Low-Literacy Populations
4-6 weeks | Information design
The scenario: A parent with limited English proficiency is trying to buy healthy food for their child. The nutrition label is a wall of numbers, percentages, and terms like “partially hydrogenated.” They buy what they can afford and hope for the best.
What you actually build: Interview people with limited health literacy about how they make food decisions. Study existing label designs from around the world (some countries use traffic-light systems, front-of-pack icons, etc.). Design and test an alternative label format that communicates key information without requiring scientific literacy.
Why this is real: Nutrition labels were designed by scientists for scientists. The people who need them most (low-income, low-literacy populations) cannot use them. This is a failure of information design.
Skills practiced: Health literacy research, information design, cross-cultural testing, regulatory awareness.
25. School Lunch Program Feedback System
4-6 weeks | Constraint-heavy
The scenario: Students at a local school hate the lunches. The kitchen knows this because 40% of the food comes back uneaten. But they do not know what students would actually eat, and “pizza every day” is not an option because of nutrition requirements.
What you actually build: Interview students, kitchen staff, nutritionists, and school administrators. Design a feedback system that gives the kitchen actionable data about preferences while working within nutrition and budget constraints. Test it for a week.
Why this is real: School nutrition programs are designed top-down. Students have zero input. The result is massive waste and poor nutrition outcomes. A simple feedback loop could fix both problems.
Skills practiced: Multi-stakeholder research, constraint-based design, feedback system design, balancing competing needs.
What These 25 Projects Actually Teach You
Every project in this list practices the same core set of design thinking skills, but each domain pushes different muscles harder. Here is how the skills break down across all 25 projects.
How many of the 25 projects build each core design thinking skill.
User research and empathy show up in every single project because that is the foundation. You cannot skip it. Problem framing appears in all 25 because every project starts with defining the right problem. The other skills, rapid prototyping, systems thinking, collaboration, and data-informed decisions, show up in most projects but get emphasized differently depending on the domain.
| The skill employers actually hire for It is not the prototype. It is the process.When you present a design thinking project in a job interview, employers do not care how polished your prototype was. They care about how you identified the problem, how you talked to users, what you learned that surprised you, and how you iterated based on feedback. The process is the portfolio piece. Document everything. |
How to Pick Your Project
Pick a problem you can access. The best project is one where you can talk to real users within a week. If the problem requires flying to another country or getting hospital ethics board approval, save it for later. Start with what is around you.
Pick a problem you care about. You are going to spend 4-8 weeks on this. If the problem does not make you a little angry or a little sad, you will lose motivation by week three.
Pick a problem, not a solution. If your project pitch starts with “we want to build an app that…”, you have already skipped empathy and define. Start with “we noticed that [these people] struggle with [this thing].” The solution comes later.
Start smaller than you think. A well-researched, beautifully framed problem with a rough prototype beats a polished app that solves the wrong problem every single time.
Go Find a Problem Worth Solving
The best engineering work does not start in a lab. It starts in a conversation. It starts when you sit across from someone and say, “Tell me about your day,” and then you actually listen. Design thinking is the framework that turns that conversation into something buildable, testable, and real.
These 25 problems are starting points, not prescriptions. The most interesting version of each one is the version you discover when you go talk to actual people and find out what the problem really looks like from their side.
Pick one. Go talk to five people this week. Write down what surprises you. That is the first step. Everything else follows from there.