by CAPRA J'NEVA
AI / ML - 15 years
Understanding what lies inside AI/ML models is key to everything from data acquisition to rapid experimentation to building AI-fueled products. My 14 year history working with the tools of AI/ML across many roles helps me innovate in the space.
The Team
Varied
Timeframe
14 years
My Role
Design Lead, Product Manager, Experience Architect
Impact
14-18% of 3.8B
Phyn market share and size
Phyn - for Belkin
In 2013-12 I designed the training routine for Phyn leak detection system. The data scientists had an algorithm that could measure voltage drop and resonance to predict which plumbing fixtures were in use. But they needed more data to make their program accurate enough for consumer use.
Working with data scientists on the team, I designed an app that would help people set up their devices, and in the process - gather and label the training data needed. The problem is - people think and say that they'll do the best job possible during setup - but my observations showed that in fact they took shortcuts, forgot fixtures, and would actually make the data science muddled.
Firstly, people forgot the fixtures they had in their house. Paper and pen prototypes helped me get to the bottom of this problem quickly. As a contractor, I didn't have time to run bloated studies, so I did scrappy, fast customer feedback. Rather than complicate things, I had the app lead them through their house, add the fixtures they saw, then turn them on and off in a routine that allowed white-space between each usage.
But reliably, no one waited for their toilet to flush to move on. 100% of people did not wait. "That's good enough, they'd say. But it wasn't for the data sicence - we needed a zero point between fixtures. Adding a 30 second timer any time a toilet was flushed, helped us get the data right.
On this job, I learned about and accounted for model drift, and reinforcement learning. I began to understand the elements of AI and how they relate to design, going on to teach my methods to others for the next decade.
Eventually, data collected, Phyn was able to release a later version that is plug-and-play for an effortless consumer experience.
My role: Principal Designer
The team: 1 product manager, approximately 30 engineers, including software and hardware engineers.
The company: Belkin
Timeline: 1 year
Knowledge service
Ontologies drove the voice-first revolution in AI experiences, enabling tools like Siri and Alexa. I built Autodesk's knowledge service in collaboration with our research team, to lay the foundations for going beyond insights on our data, into true action, automation and intelligence.
The situation: I was working on the Manufacturing Data Service - initially trying to wrangle Autodesk's 27 material databases into a cogent system.
When looking for a solution for the perpetual problem of translating mismatched data schemas across products, I met Hyunmin Cheong, an ontology enthusiast on our research team. One fundamental problem we faced over and over, was that the people who can write a schema, are not the people with the expertise to know what should be in a schema that describes real things in the physical world. Ontologies help solve this problem, by providing a graph interface that physical experts could use to describe their worlds - this knowledge could then surface in a schema editor for developers to use.
But my goal was bigger - lay the foundations to bring Autodesk into the age of AI. I was about 10 years ahead of my time. We still needed to build more data basics, before bringing back the Knowledge Service, which happened in 2026.
My role: Product Manager
The team: 2 Engineers, 1 ML/AI intern
The company: Autodesk
Timeline: 6 months
Read more in Manufacturing Data Service >
AI/ML Platform
I put together and led facilitation for a vision workshop for our AI platform team, with 40 attendees including those building the platform, and the AI/ML teams who would use it to condition their data, develop and deploy models. Our goal was to identify all the capabilities the platform needed, create a northstar storyboard and identify which capabilities should be shipped for an MVP.
I created the facilitation method we used to enable teams to see beyond the near term, and define a product vision that would be durable across their end goal, to avoid short-sighted hacks that they would pay for later through refactoring.
The conversation was so good that the workshop host told me he had learned more in 3 days than in an entire year of working on the project. Defining the minimum lovable product was a breeze, because of the insights we gained from the discussions. Our storyboard formed the basis of product management's stories for the next 2 years.
For more AI projects see:
Collaborative AI
MCP Servers