[4eyes] FW: OpenCV.ai hiring

[Matthew Turk]

FYI, interesting email from Gary Bradski (founder of OpenCV, currently with Magic Leap) below. 

 

From: Gary Bradski [mailto:garybradski at gmail.com] 
Sent: Wednesday, May 11, 2016 3:41 AM
To: Ethan Rublee <rublee at opencv.ai>
Subject: OpenCV.ai hiring

 

All (many Bcc'd) 

 

EXEC SUM:

We are trying to hire up a small team of super coders in vision to make the "out of the box" experience of software+camera+compute (smart cameras) just work to accelerate the field by making sensing completely parameterisable (# of cameras, timing relationships, exposure relationships, lens or lenses, all simple to configure and work with). We want to step vision up another notch for everyone. But we need damn good people/coders, just a few, to do this. Send them to us please! We have lots of hardware partners to complement the effort!

 

WHY:

I've decided to form a for profit vision company around, and in support of, OpenCV (library will still be free and open). The reason is to further the mission of OpenCV to bring visual intelligence to the world by bringing an inexpensive professional development to people along with a pipeline to product for us and everyone else. OpenCV will remain free and open with better support, the development platform will itself be an open spec.

 

WHO:

Ethan Rublee, CC'd will be CEO. He was a co-founder of Industrial Perception, the robotics company we sold to Google. He's leaving Google X to run OpenCV.ai, and is, what is known colloquially as, a super producer.

 

Gary Bradski is co-founder, a funder, advisor, board member.

 

We have a guy, "Jack" in China. You don't know him. He's a vision business savvy guy there.

 

BACKGROUND:

Computer vision cameras, drivers and SW involve too much effort to get to an effective baseline. The cameras and lenses are too limiting. Time to change that.

 

GOAL:

So, OpenCV.ai is going to open with teams in the US and China with the goal of becoming the computer vision platform of choice on which we and others will build products.

 

WHAT:

*	Build an inclusive kernel, Linux+OpenCV+embedded deep+eigen+opencl+opengl, wedded and optimized with "OpenCV Certified" hardware makers (cameras + compute = smart cameras) multi-smart camera development system (for $ale, but cheap), that

*	Provides great "out of the box" vision experience.

*	Just works, no other dependencies, great API
*	calibrates, syncs, phases with other cameras
*	does geometric vision well
*	provides lots of baseline functionality
*	easily ingestes and runs deep nets trained by your favorite deep net engines

*	Drives 6 or more cameras per board, dozens of such boards can be put together and will sync, phase control as they are linked together.

*	Experiment, print

*	Smart cameras, linked together are already a super computer.
*	Develop straight on the camera/camera array. Chrome cast the results to your "computer". When you are done, it is already running on your target platform
*	Via "OpenCV Certified" relationships with hardware makers, you can just "print off" one or 1000's of products related to what you did on the development boards.
*	Market place of licensable best of class other algorithms, deep nets.

*	Make working with depth cameras, configurations of many cameras and especially working with light fields trivial/common. All those lost lightfield papers of the 90s will come back!
*	Students will do more interesting research when it is trivial to deploy more interesting vision systems.
*	OpenCV.ai together with others will build capabilities and products on top of it all.

QUAL:

This isn't a call to see which student, yet to find themselves, might procure employment. Ethan needs a few people around him who can code and hardware hack at his level. His level is high. Extremely so.

 

Let Ethan or myself know.

 

We, and vision, thank you!

 

Gary

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cs.ucsb.edu/pipermail/ilab-users/attachments/20160511/0a9e1790/attachment.html>