Keith Carlson | Studencki Festiwal Informatyczny

16th SFI Academic IT Festival

- October 2020 |
16 edition 2020

Keith Carlson

Come from: Chicago, USA
SFI Edition: 16

About Keith

Keith Carlson is the chief technology officer of Relativity. As head of technology and architecture strategy, Keith oversees the Engineering Delivery, Engineering Operations, and Production Engineering functions. In his career, Keith has been a co-founder, programmer, team lead, principal engineer and chief technology officer. He most recently spent 7 plus years at Amazon Web Services where he developed the first cloud fraud prevention and detection organizations and grew it to where it was evaluating all of the activity at AWS to identify fraudulent behavior. Keith is excited to apply his extensive developer experience to help mature RelativityOne’s SaaS delivery model and expand Relativity’s reach into the unstructured data realm.


What do you think is the most satisfying part of your work / technology?

The most satisfying part of my work is building things that make someone’s life easier. It’s satisfying when I speak with customers of our software and they say, “This is really cool, now we can do [something] that we could never do before”. Being able to take something that is hard or seemingly impossible with current technology and making it possible is the most satisfying part of my work.

When will AI “take over” the work of programmers?

I have been in the tech industry for 30 years, and I often hear about technologies “taking over” work. In reality, what happens is that work shifts to higher value things, or the requirements for a given solution increase. For example, in metrics gathering and reporting, new tools have not taken over, rather the requirements for what is gathered and analyzed increase. So when I began my career, many decisions were made based on intuition, but now they require metrics and data. For programming, I do believe that coding such things as database connections or data input screens are approaching automation, but I believe that will only lead to more requirements for the core of the software. AI/ML/Deep Learning work well with lots of data about how a particular pattern has behaved in the past and when it has been repeated frequently. However, it struggles when there is not a lot of data or the pattern has not been repeated many times. The areas of programming that are commonly repeated will get automated, and the more specialized work of the uniqueness of the software will simply become more sophisticated.

What would you like to do technology wise, as if you had unlimited time for it?

It’s always been a mystery to me how we can build the most amazing things using programming languages and editors that have not materially changed in 40 years. If I could do anything in tech, it would be to fundamentally re-think programming languages and integrated development environments. When I look at a new 3D gaming User Interface/User Experience, I am amazed at the sophistication and how real it feels to interact with the virtual world. Then I sit down to write a program and use a language that is decades old and have a tool that opens multiple text files for editing. It is counterintuitive that we have not significantly enhanced the tools we use to write programs