If you haven’t heard Ramona Pierson’s story, you’re missing out on a great source of inspiration. It involves a broken home, a teenage math phenom, a stint in military intelligence, and a career in public education and artificial intelligence systems, including her latest startup endeavor called Declara. The most-significant event might have been a horrific car accident at 22 that left her all but dead and in a coma for 18 months. (The San Jose Mercury News published a profile of Pierson in December that covers many of the details.)
Pierson came on the Structure Show podcast this week to talk about Declara and its approach to building professional learning platforms that can scale across large national institutions. It is an interesting discussion (and you can listen to the whole thing below, which includes Barb Darrow and I talking about the intersection of big data and sports), but the interview is perhaps best summarized by Pierson’s account of finding herself left to rehabilitate in a senior citizen’s home after nearly two years in and out of VA hospitals after her car accident.
Although her companies have focused on machine learning technologies (she’ll dive deeper into that at our Structure Data conference next week, March 19-20, in New York.), Pierson’s experience relearning how to walk, talk and see has informed her approach to what she builds:
“When I landed at the senior citizens’ home, imagine somebody showing up and they’re 68 pounds, bald, not able to walk, talk. I was completely blind at that time, and so they probably thought E.T. just arrived. At that point they had an emergency session and they — imagine a hundred grandparents and they look at you and they think, ‘OK, we need to figure out what to do with this person.’
“And I was their sole point of interest. They had nothing better to do than to focus on me. So my days and nights were filled with these assessing me and testing out ways to teach me new things. And so because I was blind, and none of these people were specialists with the blind, they kept testing things out.
“What I learned through that experience is to assess people, to try out different modalities in which people may learn best and different pathways. Because all of us may learn things in different order. So when I look at the precursors and post-cursors of learning a certain competency, we may have to adjust those based on how people learn best. … So many of these programs that are out there are making assumptions that everyone learns the same thing in the same way, but that isn’t true.”
With Declara, she explained,
“We really try to personalize around preferences and validate those preferences. So, if someone says I’d rather learn through video or rich media, but yet we actually start to validate the system going, ‘Oh, they actually do best when they’re trying to help others. When they’re an expert in helping others, they’re actually learning faster.’ So we’ll take in the preferences and things that people declare about themselves — thus [calling the company] Declara — but then we also validate with our programs.”
The approach isn’t entirely uncommon — many systems built on top of machine learning and artificial intelligence try to learn from user behavior — but Pierson says Declara’s methods are unique. To the extent they are, it’s probably because Pierson’s experiences are unique. She’s not trying to apply the lessons she learned at a large web company and scale them down to a new field or application. Rather, she’s trying to apply the lessons she learned about learning and scale them up to help thousands of people at a time.