Prototyped a machine learning algorithm to recommend recipes based on user input with a 74% case success rate
Finding flavor in home kitchens
I led our team through a LEAN product process to find product-market fit with the least operational waste.
Conducted controlled product testing, which indicated increases in confidence, happiness, understanding, and satisfaction when using the product
Designed and led multiple sprint cycles as a part of a rapid iteration strategy
- Mobile Apps
- User Research
- Data Analysis
- Product Management
- Product Discovery
- User Experience Design
- Market Research
- Web Apps
- People Management
Fighting food waste by finding what you love
Food waste is a growing concern in the United States, costing the average American approximately $1,800 yearly. This cost is a direct result of poor planning and low confidence in home kitchens. Ripe fights food waste by using AI to help users explore new recipes, plan their meals, and gain confidence in the kitchen.
Using LEAN to drive product operations
As the Project Lead for my group, I guided our team through LEAN UX techniques. Rather than spending 50% of our time researching, 30% designing, and 20% testing, we created product hypothesis models early. We then ideated and user-tested continuously and iteratively using rapid prototyping techniques to find the ideal solution for our users as quickly as possible. I planned and led multiple sprint cycles with clear phases and goals to encourage iteration, reflection, and minimum-viable product mindsets.
Simplifying complex technology into low-cost prototypes
Ripe is built on two key AI platforms: a voice-based platform that guides users through a recipe and a machine-learning platform that suggests recipes based on user input and feedback. In a final product, these platforms would be expensive and time-consuming to build -- even more so if we got them wrong. To limit the operational waste put into these algorithms, we used gorilla testing methods such as spreadsheets and Wizard of Oz to mimic the intended products as closely as possible without taking on the cost of building them.
Creating quantitative product measurements and outcomes
I designed, led, and analyzed two key testing phases as a part of our Agile mindset. The first compared a control group's emotional factors before and after cooking to an experiment group, the latter of which used a prototyped AI assistant. This test resulted in measurable, statistically significant outcomes: after cooking with Ripe, our users were more confident, happier, and calmer than those who cooked without, and they reported a higher understanding of the ingredients they cooked with. Second, we tested our machine learning algorithm's ability to suggest recipes, measuring our ability to identify the traits a user looks for when selecting a recipe. This test returned more favorable results with over 74% of cases having more than two-thirds of suggested recipes accepted.