Solving the mystery of sleep


AMLG: Welcome back to the pod. I’m excited to be here with Dr. Assaf Glazer. He is the co-founder and CEO of Nanit a leading human analytics company that uses computer vision to help parents navigate their child’s sleep..Essentially it’s a baby data collector that every sleep-deprived geek parent has dreamed of.

A little background on Assaf: He got his Ph.D. at the Technion in Israel and was previously at Applied Materials as well as Wales where he worked on solutions for missile defense systems. Nanit was born here in New York at Cornell Tech [disclosure — RRE is a long-standing investor in the company.] Welcome Assaf it’s great to have you. 

AG: Thank you for having me.

AMLG: I’ve got a stat here, that on average parents lose 44 days of sleep during the first year of their baby’s life and nearly 3 in 10 babies have problems sleeping at night. Those numbers sum up the nature of what you’re trying to solve, but can you lay out how you identified this problem and started the company?

AG: It started for me as a parent. You have your baby, you arrive home and you see that your life has changed. Pretty quickly you understand what your number one concern is — sleep. You’re tired, you’re sleep deprived. You wake up during the night and do everything necessary to go back to sleep. You’re going to Google and going to friends. This is where Nanit comes in. We are giving you the information that will allow you to make better decisions for your child. Six years ago I had my first child, Udi. He was born when I was at the Technion. I’m a computer vision guy. Before I was at the Technion I worked at Applied Materials in the semiconductor industry, on a camera that you put above the silicon slices, to see them from a bird’s eye perspective.

AMLG: So you were doing computer vision for chip manufacturing — on the assembly lines, you’d look for errors in the chips?

AG: Yes. And when my son was born I said, OK let’s do process control for my baby.

AMLG: As if the baby was on an assembly line like a chip, just run some computer vision on it. 

AG: Yeah. So I wrote a paper on background subtraction algorithms — how to find a foreground object differentiated from the background — and applied those algorithms to my baby. I went to my advisors at the Technion and told them, you know, I’ve found that my baby is moving 134 times on average at night. But what can you do with that? I was looking at this data and I said sleep, sleep is what we need to solve here. I went to sleeping labs to try to understand sleep science. Then I moved as a postdoc to Cornell University where I joined the Runway Program, which aims to commercialize science.