Brian Brackeen of Kairos Facial Recognition

Despite not having graduated college, Brian has worked for big corporations including Comcast, ADP, IBM and Apple before founding the facial recognition company Kairos. Today on the EO Podcast, Brian discusses the how Kairos was started, and explains how A.I. and facial recognition works for everything from movie screenings to cruise ships. Tune-in to learn how Brian handled the evolution of Kairos, his worries about the future of A.I., and his beliefs on privacy regulation in the tech world. Brian also explains his take on how he wasn’t “black enough” for Inc magazine.

Time Stamped Show Notes:

01:07 – Brian is the founder and CEO of Kairos, an A.I. company focused on facial recognition, in 2012 01:28 – Selected by Wall Street Journal in 2013 as one of the top 25 start-ups, was previously a Senior PM for Apple, a senior managing consultant for IBM, and also worked for ADP and Comcast 02:17 – Also lectures on entrepreneurship, code, A.I., computer intelligence, digital economy, and participates in mentorship programs like “Girls who Code,” “Black Girls Code,” and through Miami’s school district 02:56 – What is the motivation behind getting so involved? 03:08 – Grandfather was Baptist preacher and was always giving back, this was engrained from young age 03:30 – Adopted at 6-8 months and was given great childhood so wants to show the same generosity, had Amish foster parents – ironic considering his career path 05:48 – He got poor grades but then went to Penn State...how? 6:25 – Got into a commuter version of Penn State but never graduated 06:42 – Ended up in Miami twice 07:50 – Was picked up by Apple despite not having a college degree because he could code and was senior level at the previous companies he worked for 08:20 – Believes higher education is not always necessary; it depends on degree 09:20 – Apple story: MobileMe had poor user experience, wasn’t a great product, and Steve held team meeting to ask everyone what MobileMe means to them. 11:40 – After hearing teams positive responses, Steve screams, “Then why the fuck doesn’t it do any of that?” 12:23 – Elevator story 13:00 – He was told to leave elevator if Steve gets in elevator with him because he was new and didn’t have a prepared answer to the inevitable, “What do you do here?” 14:16 – Kairos was started as a punch card system to combat “buddy-punching,” but quickly realized it was more of a facial recognition company 15:15 – He tells young entrepreneurs that whatever you are doing year 1 will be very different year 2 or 3 15:40 – You need to be able to see the signs of change and move the company until you find the perfect product market fit 16:03 – At first facial recognition was the biggest challenge, but once they got it the entire company changed; the product was strongly differentiated, increased in value, and providing it to enterprising companies changed everything 16:35 – 3 years in, the company pivoted to an entire focus on facial recognition for enterprise customers 16:45 – Kairos can be consumed in 2 ways: As an API and STKs/on premise solutions 18:30 – Kairos can focus on motion, age, gender, sentiment, ethnicity (and mixed ethnicity) 19:10 – Kairos can learn everything about a face and can tell you the % of ethnic groups in the face 19:40 – It wasn’t coded that way; A.I. learned the facial groups and human genealogy over time by itself 20:10 – A.I. is so powerful because it can find insights that we otherwise couldn’t see 20:45 – Scary/worrisome things about A.I. 21:08 – Implicit biases: Biases can be trained into the algorithm by mistake and it’ll be repeated over and over again at scale, this is worrisome 21:35 – Example of implicit biases: faucet sensor that was only trained with Asian or white hand, so the faucet wouldn’t work with African American hands 22:25 – Statistics on accuracy of Kairos’ facial recognition software 22:50 – Software can find 1 person in 1 billion in 1/30th of a second and they are 99.8% sure that person is who they think they are (based on photos) 23:30 – Any camera, any image, 70 pixels between the eyes (iPhone is 1000 pixels across) 24:15– Use cases 24:25 – Used by movie studios for audience emotion recognition and to determine the likes/dislikes of specific demographics 25:00 – Frictionless checkout: Being able to walk onto cruise ship without stopping because data is already in system and face is recognized upon entrance 25:20 – The retail experience: Walking in, grabbing what you want, looking into a camera, and card is automatically charged 25:40 – New uses: Automotive and health care industries: Doctor pulls up medical records with facial recognition and cars with emotion recognition to drive to users liking 26:45 – How can a child who ages over time still be recognized? 26:50 – Post-puberty: Certain spaces and bone structures won’t change with age or weight 27:27 – Issues with regulation, morality, privacy – Where is the line drawn? 28:10 – They are one of only companies that work closely with privacy regulation, they believe there should be consent or a relationship between company and scanned person in appropriate spaces 28:40 – Others in industry are anti-regulation (including Facebook) 29:28 –Has received $8.1 million in funding...how? 29:40 – Put in $250,000 himself, and the other $250,000 came from investors – this was in angel/idea phase 30:29 – The relationships he made were key in making it possible 31:20 – Had early customers and traction a year in with good projections 31:40 – $1.35 million in priced round, evaluation was $4 or $5 million 31:50 – Led by New World Angels with VCs 32:00 – He is a majority shareholder and the angel group and VCs are still involved and have reinvested 32:30 – 40% of company given away in equity; they have an employee option pool so some employees have equity 33:00 – Had a co-founder in the beginning, but as they got bigger she chose to move on 34:16 – “How black are you?” 98% 34:45 – He wrote “I’m not black enough for Inc Magazine” article 35:00 – He’s actually 15% Welch 35:15 – Inc magazine was following him around and interviewing people around him to do an article on him and African American entrepreneurship 35:56 – He got an email stating that the editor killed the story because he doesn’t represent the “archetype of a black entrepreneur” 36:20 – The person they wrote about instead had been to prison and had a rags-to-riches story 36:30 – His problem with their decision is that in their minds he and his background is not what an African American’s should be; when in fact the average African American is middle class 37:11 – Believes that we need to get past feeling the need to paint a negative narrative for African Americans   “Boxers or Briefs” segment The Matrix or I, Robot– The Matrix IBM or Apple – Apple End of human kind via A.I. or Meteor – Meteor Beyoncé or Shakira – Both Michael Buble or Michael Jackson – Michael Jackson Coffee or Tea – Tea Netflix & Chill or Dance Party – Netflix & Chill Snapchat or Text – Snapchat Email or Slack – Slack IPhone or Android – “PLEASE!” iPhone Dog or cat – Cat Miami heat or Philly snow with a side of cheesesteak – Philly snow...because of the cheesesteak 40:19 – Ethnicity test anyone can take: Website 40:45 – Test is free to grow awareness of what facial recognition can do and shows them what is possible, and system can get smarter from the tests 41:30 – Reach Brian via Twitter Dave closes the episode and encourages you to visit his website

3 Key Points:

A college degree isn’t always necessary depending on the field and one’s ambition and willingness to self-teach. It is important to recognize your company’s necessary changes and evolution so you can find your perfect product market fit. Mentorship and transparency are important in the A.I. field to show people what is possible, open up the relationship between the company and the people involved, and understand when regulation is necessary.

Resources Mentioned:

Entrepreneur's Organization – The EO Network Kairos – Brian’s company New World Angels – Organization of angel investors “I’m not Black Enough for Inc Magazine” article written by Brian

Credits:

Show Notes provided by Melissa Valder