Underdog Founders - Panos, CEO of Zenus Inc.
Everyone loves to watch underdogs win. Panos is as good as they come.
Is it possible to roll out facial recognition for crowds without waking up in 1984?
Panos’ life path changed early, back when he was in Greece doing his bachelor’s in Statistics and Math.
One of his professors saw something in him and pulled him aside.
He asked point-blank:
“You’re doing really well here - what are you planning to do with your life?”
As most people in their early twenties, the truth was simple:
He had no idea.
Or more accurately, Panos had a tiny speck of a fuzzy idea:
He didn’t know what, how, or where.
But he knew he wanted it to have a positive impact.
He met a professor from the University of Houston (UoH) who convinced him to do his PhD there, and gave him a full scholarship.
Fast forward a bit and UoH filed a patent based on Panos’ work, so he found himself at NSFI Corp - National Science Foundation Innovation Corp.
They fund researchers working on interesting tech, to allow them the freedom to explore the commercialization potential of the patents they’ve been working on.
Think university research meets Business Model Canvas.
One day while talking to peers, someone in his group had an idea:
Could we use this research to do facial recognition in crowds?
Maybe it would be useful for events that want to check-in lots of people quickly?
The idea got some heads nodding, they raised a tiny amount of money, and decided to focus on the topic without knowing much about the industry.
They had no idea back then, but a bunch of Big Tech companies already tried and failed spectacularly and expensively.
The value of tech was super clear, but so were the struggles:
1 - Privacy concerns.
2 - Capacity to upload large spikes of data in crowded areas with shaky internet connections.
3 - Low margins.
Facial recognition would never work with 10 or 50 cameras recording 4k video that somehow had to get uploaded to the cloud all at the same time.
Even if you could transfer all those GB per camera, then you had to process it - everything was too expensive in this model.
Seeing how much money was spent brute forcing this to get it right, motivated Panos and the team to prove the approach was fundamentally wrong.
They got to work on a solution focused on edge computing - capturing and processing data inside the building, replacing the bandwidth problem with a hardware one.
Plus, it was fascinating to see if facial recognition could be done in an ethical way.
The focus was on facial analysis, to compute crowd statistics without identification. Nobody is ever identified or matched with a database.
Tech companies should bear some responsibility for the use cases they unlock - it’s not just the end user that should be accountable.
“We focus a lot on the real world implications and consequences of our tech, that we jokingly say we are building the anti-metaverse - all we care about is how tech impacts the real world.”
So, Panos starts working on this in 2020.
Living off of your research in the US meant surviving on a net income of 25k.
Doable in Southern Europe, but in the United (much more expensive) States, he was scraping the bottom of the barrel.
“I could be making above $200K comfortably in Big Tech. Yet there I was, maxing out credit cards, supporting my girlfriend when she got sick while building a pre-revenue startup.”
Their product also happens to be super horizontal - hotels, airports, campus, event spaces, any area where large crowds will gather.
So you can imagine CV19 basically pulled the plug on revenue.
I’m talking all the way down to −273.15 Celsius:
Absolute zero.
“We’re only alive because we were able to raise capital through our early angel investors and equity crowdfunding. Hundreds of people believed in our vision.”
He had moments where he would literally eat less because the belt was that tight.
Working days and days until 1am for 5 or 6 times less than what you could make in Big Tech…
You only do it without giving up if you’re in it for the right reasons.
Today, Zenus is growing at 300% YoY for two years in a row now.
Their clients get so much value it even taught Panos to love sales - because he already knows people are gonna love what they built, and it makes him happy.
Team went from 4 people to 25 by the end of this year.
And the product, for all the headaches it gave during development, is now performing above expectations.
“A fond memory we have comes from a client test during a packed live event. The product was handling the data smoothly even when the lights of the venue came down to focus on the main stage.”
“We thought there was no way Zenus would keep tracking audience sentiment in such low light, but there it was - simply chugging along without a care in the world.”
It’s a funny feeling when you build something and then your creation does things you didn’t expect. I bet good ol Sam (Altman) can relate.
From their favorite testimonials, 4 say the same thing:
“this is liquid gold”.
Proof that PMF can also be found through words.
If you know someone that manages top tier events, forward this.
If you’re keen to see how AI and facial analysis can be used for good in an ethical way, follow Panos here on LinkedIn.
And if you love our content, consider supporting us because we do this for free: go premium for 26 cents/day.
Underdog Founders 13 - Panos Moutafis from Zenus.
#founderstories #edgecomputing #facialrecognition