Case study
EmberEye: fire prevention through thermal imaging & real time danger alerts
Each year, 150,000+ U.S. kitchen fires start from unattended cooking. Smoke detectors are often disabled due to false alarms, leaving families unprotected. As the sole designer on a 10-person engineering team, I created EmberEye: a thermal-sensing safety device and companion app that gives home cooks live alerts and stove footage—stopping fires before they start.
Date
9 months (2023-24)
Project type
UCSB capstone project sponsored by Teledyne FLIR
Role
Sole product designer (UX, research, branding)
Team
5 EE, 4 ME, 1 iOS Dev, 2 FLIR advisors
Problem
Kitchen fires cause 150k+ U.S. incidents yearly. Smoke detectors often fail users.
Solution
EmberEye delivers real-time alerts and stove monitoring through a thermal-sensing device and iOS app.
Impact
User
3/3 users successfully completed setup with prototype
Business
Projected prevention of 50k+ fires/year; ~$18M revenue potential at 25% adoption.
Recognition
1st Place Capstone Award for cross-disciplinary innovation
Key Contributions
UX Advocacy
Shifted team from “feature-dump” mindset to outcome-driven UX strategy
Vision & Buy-In
Pitched app concept, secured buy-in, recruited iOS dev.
Led End-to-End Design 0 → 1
Designed onboarding + alert flows through iterative prototyping, research, and testing.
Engineering Collaboration
Partnered with engineers to define data thresholds and safety priorities.
My team and I after our presentation at the FLIR campus in Santa Barbara :)
[Solutions at a glance]
Designing peace of mind in the kitchen
EmberEye focuses on the moments that matter most—easy setup, clear stove monitoring, and real-time danger alerts—turning complex thermal data into simple peace of mind.
Background
Why existing smoke detectors aren’t enough
One-third of Americans disable their smoke detectors because of false alarms or battery issues, leaving homes unprotected. Kitchen fires remain the leading cause of household fires, yet traditional detectors only react after smoke appears. The challenge: design a smarter, proactive solution that prevents danger before it escalates.
Here are some fire-starting origins:
Unattended cooking situations before they escalate
Overheating oil scenarios where smoke precedes ignition
Slow-developing grease fires that emit smoke before flames appear
Research & Prioritization
Designing within real-world constraints
We had a $10,000 hardware budget and limited time. To stay realistic, I scoped EmberEye for households with modern kitchens willing to invest in premium safety (around a $250+ price point). This constraint pushed me to focus the app design on core value: trustworthy alerts and a simple, anxiety-reducing interface.
Who we designed for
Without access to direct users, I created early-stage archetypes using secondary research and observational profiling. These insights helped us keep user needs at the center and drove design priorities: real-time notifications, low-friction setup, and clear visual feedback.
The Family-Oriented User
juggling caregiving and cooking, need immediate alerts for safety
The Average User
living alone or multitasking, want convenience and peace of mind
The Neuro-divergent User
benefit from reminders and live stove footage to reduce anxiety
Learning from the market
I audited five physical-digital products (such as connected appliances and security devices) to understand how they onboard users,
Audit of other competitors
Key takeaways
Connection & Pairing
Use QR codes or button presses for easy pairing. Pair tactile and visual indicators to build trust that devices are working.
Configuration
Simple guiding questions during onboarding helps the system understand user’s unique home environment
Instructions
Add progressive disclosure and animated guides to prevent overwhelm. Keep instructions concise and break them into small, visualized steps.
Prioritizing what really matters
Working with engineers, I facilitated a workshop to rank which kitchen events needed critical alerts versus simple notifications.
Examples included:
Active fire → blaring alarm plus push notification.
Boiling over pot → push notification only.
Left burner on → low-level notification.
This exercise aligned the team around safety-first prioritization and kept us from overloading users with unnecessary alarms.
Worksheets the users filled out
Rapid prototyping & testing
Prototyping fast to align stakeholders and validate setup flow
With limited time and no prior app context, I created a high-fidelity prototype of the onboarding flow. This let me:
Communicate clearly to non-design faculty and engineering teammates
Quickly test usability with early users and identify pain points
Notes from faculty presentations and user sessions revealed three key issues: information overload, weak visual hierarchy, and lack of visual guidance. These insights directly shaped later refinements.
Notes on V1 prototype
What I learned from early user testing
Information Overload
3/4 participants (1 neurodivergent) found that the instructions and recommendations were too wordy, users did not know which parts are the most crucial to focus on.
Lack of Visual Hierarchy
2/4 participants found it difficult to tell which part of the screen users should focus on.
Lack of Visual Guidance
1/4 participants were unsure where exactly to mount their stove with only textual instructions. Some steps were confusing.
HMW
Simplify the device set up process and reduce users' cognitive load to reduce churn?
Refining the Onboarding Design
Targeted instructions for different kitchen setups
Early testing revealed that a one-size-fits-all setup flow overwhelmed users. To reduce cognitive load, I split the onboarding into two tailored paths—magnetic vs. non-magnetic range hoods.
This allowed me to:
Simplify choices by showing only the steps relevant to each user
Highlight critical actions without clutter from secondary info
Streamline setup so users felt confident mounting the device correctly
By personalizing onboarding to real kitchen environments, I made instructions easier to follow and significantly reduced setup friction.
Before
After
Before: one flow for all users created confusion. After: branching flow reduced overload.
Reducing clutter with secondary hints
Users were overwhelmed when every detail appeared at once. I moved secondary instructions into expandable hint cards, letting users access them only when needed. This reduced visual noise while still supporting those who wanted extra guidance.
Moving extra details into hints kept primary instructions clear and scannable.
Guiding setup with visuals and animations
Text-only instructions left users uncertain about mounting and setup. I added:
Shutter animations to confirm steps were completed
Burner listings so users knew exactly what to drag in
Lightweight tutorials with visual cues instead of long text
These changes made setup more intuitive and directly addressed hierarchy and guidance issues.
Visual Feedback
Shutter animation indicates when a screenshot is taken. This visual feedback reduces cognitive load by clearly signaling that a step was completed, ensuring users understood the process without confusion.
Listing the burners
Listing all available burners upfront gives users clear visual feedback on how many they need to drag into the frame.
Animated tutorials
Animated tutorials with visual cues allows them to succeed without needing to read instructions too carefully.
Building trust through personalization
The onboarding required several steps, and some users doubted whether the process was worth it. I designed a final “personalization” screen that framed the setup as tailoring the system to their unique kitchen.
According to Peter, the founder of BFM, "if we think something if personalized, we believe it'll be better suited to our needs."
This moment built anticipation, reinforced trust, and reframed the lengthy flow as a payoff: “This product is designed for me.”
A personalized end screen built confidence and trust in the product.
Why it all mattered
By simplifying information, guiding with visuals, and closing with personalization, onboarding shifted from overwhelming to empowering. Users felt confident they had installed the device correctly, a critical step in ensuring long-term adoption and reducing churn.
In-app Experience
From alerts to peace of mind: designing the EmberEye app
With setup complete, the next challenge was ensuring EmberEye could warn users clearly, reduce anxiety, and provide confidence their kitchen was safe. My early prototypes were feature-heavy, but testing revealed that simplification and clarity—not more features—were what made the app intuitive and trustworthy.
HMW
Effectively alert users and help reduce user stress regarding potential dangers in the kitchen?
Evolving from cluttered to clear
I began with a concept sketch full of imagined features, but usability tests showed the need for focus. Through multiple iterations, I simplified the dashboard, adopted native iOS patterns, and created a layout that balanced clarity for users with feasibility for the developer.
From feature overload to a streamlined, native iOS-style dashboard.
Alerting Users
Clear alerts for urgent moments to make danger impossible to miss
Users needed to instantly recognize danger states. I designed alert cards that shifted color, background, and hierarchy based on urgency, making risks like unattended cooking or boiling over impossible to ignore.
Dashboard alert state
Different states of the alert card.
Burner monitoring for multitasking cooks to support awareness in complex cooking
Testing showed that home cooks—especially during multi-course meals—often lost track of which burners were on. I added a burner monitor at the bottom of the dashboard to provide this awareness without overshadowing critical alerts.
Burner status at the bottom helps users stay in control without cluttering the dashboard.
Push notifications for peace of mind on the go
To reassure users outside the home, I designed push notifications that surfaced critical stove activity and linked directly to the alert dashboard. This turned EmberEye from a static device into a trusted safety companion.
Notifications keep users informed and ready to act, even when away from home.
Testing under constraint
With limited access to end users, I ran quick “black hat” tests with my engineering teammates. Using prompts and lightweight Google Forms, I gathered feedback on clarity, hierarchy, and alerts. This let me refine designs at every tech update, keeping usability top-of-mind without slowing engineering progress.
Features for a Peace of Mind
Users worried about false alarms during normal simmering or baking. To reduce alert fatigue, I designed a timer system that suppresses unnecessary warnings until cooking is complete. This aligned alerts with real cooking workflows—turning them from nagging interruptions into helpful kitchen partners.
Custom timers prevent false alerts and match real cooking behaviors.
Live stove feed for instant reassurance and reduce anxiety
The most loved feature was a live video feed of the stove. If an alert triggered, users could instantly check if it was real or false, and those prone to “Did I leave the stove on?” moments found comfort in being able to confirm remotely.
Live view reassures users, reduces anxiety, and prevents unnecessary trips home.
Impact and results
The final product prototype was successful among participants and final demo
I tested the near-final prototype with 3 participants in their kitchens to evaluate setup and usability. All participants successfully set up the device and app, and most reported reduced anxiety about kitchen safety. Feedback highlighted the product’s value for families and neurodivergent users.
3 out of 3
participants completed setup successfully
2 out of 3
participants + 5 faculty members said the product eased safety anxieties
1st Place
Our project won the excellence in multidisciplinary award!
Group pic after our award ceremony!
Reflection
How I grew and what I learned
As my first time working as the sole designer on a cross-disciplinary team, EmberEye pushed me to grow beyond design execution. Here are the three most important lessons I took away:
💡 Design is influence, not just execution
I learned how to win buy-in from teammates and faculty by framing the app as essential to the device’s value, not a “nice-to-have.” This shifted the team mindset from feature dumping to outcome-driven design.
🤝 Collaboration means speaking their language
Working closely with electrical and mechanical engineers taught me to bridge the gap between UX and technical feasibility. My mechanical engineering background helped me translate technical constraints into design opportunities instead of blockers.
🌱 Scrappy beats perfect
Without access to formal users, I got comfortable running scrappy tests with teammates and iterating quickly. This mindset taught me how to keep design moving forward even in imperfect conditions—a reality in most product teams.
Next Steps
What I’d explore with more time
Given more resources, I would expand testing to measure impact at scale and validate real-world usage.
Measure time on task to identify friction during setup.
Live testing in real kitchens to observe how users respond to alerts (Do they check live view first? Call someone? Ignore it?).
Refine live view feature — test recording duration, rewind limits, and explore AI for labeling footage.
“The app Felix created was by far the best app I've ever seen come from a capstone project… certainly the best from a usability and functionality perspective.”
Next up
Improving trust among college passengers and drivers for PoolUp, a city-to-city rideshare early stage startup!