Podcast Content Studio

Ep 190 Kirk Marple of Unstruk Data on the Unstructured Data Challenge and Lessons of a Technical Founder 

Kirk Marple is interviewed by Eric Wright on DiscoPosse. In this episode, Marple and Wright get to have a very tech centered conversation that breaks down Unstruk from the business perspective into its origins as Marple’s passionate and consistent creation through code. Marple speaks about his excitement to launch Unstruk and his desire to simply optimize workflows. 

In this episode, you will learn:

• In depth conversation about a variety of stakes related to metadata storage
• What Marple visualizes the company being able to provide for future innovations
• How being cloud based helps with scalability
• Remaining grounded in realism when it comes to fitting every customer’s needs
• Building the infrastructure of Unstruk
• More tech-specific discussion on certain aspects of building the platform like the Knowledge Graph and certain tools Marple prefers. 

Good Question & Answer 1:

Wright: 

When this is where the scale takes you to a difficult place quickly, in this world, right? when you start getting into hundreds of 1000s, and millions like, and you have to ingest that data and think about storage, and think about performance, that’s where it gets really tricky.

Marple: 

We’re in an interesting spot, where even more people aren’t understanding how to ultimately extract every possible inch of value out of content and this data. They know they need to collect it and they need to store it. So we’ve got the first phase, which is ideation  ‘Well, we’ve got this potential to capture this data, then let’s try and get insight out of the data.” so a lot of people within the last decade are doing a lot more just raw collection. Then they don’t know what to do. That’s really what we are really being opinionated about is data. Ingestion is where we’re starting. People have data sitting in s3 buckets and having a blob store to have it in, Dropbox or something like that. Getting it into a platform and indexing it, it’s been done in a sense, but it hasn’t been scaled very well- for and across all the different formats. Everything from a point cloud to a CAD drawing. That’s the first stage. Then data organization, auto organizing that data- one of our investors is a an oil and gas company and they have hundreds of 1000s of videos of the undersea floor of where their oil pipelines run. What we’ve heard is they just sit in an s3 bucket somewhere, they’re not well indexed by time and and geospace and stuff like that, or even about what’s in the videos themselves. So we really see a platform that can just get access to the data and organize it as a first step. Then it’s about exploring the data, “I want to filter by that region of the North Sea, or that region of the Pacific Ocean. And I want to look in this month in 2017.” That’s really the next part we’re pretty opinionated about, and then the visualization. 

Good Question & Answer 2:

Wright: 

With your understanding at real enterprise scale, you know, how important was that in your ability to now have a go to market with this product because this quite often is a really tough battle when you go especially the companies at the scale that are going to leverage this. 

Marple: 

it’s a really good point. I was at General Motors for a time working, that was where I first got exposed to industrial media or industrial unstructured data. I’m not even sure if they would be a customer for it because it’s a bigger ecosystem. There’s a slice of customers or companies that I’m just not that worried about like Uber,  Tesla, they’re not going to use our software, they’re gonna build it themselves, they already have tools for for stuff like that. But a notch down, there’s a huge swath of industrial companies, from manufacturing to chemical to robotics, OEMs, and things like that, that I believe can really make use of this. It’s not going to solve every problem they have, but hopefully, it’s a catalyst. Let them do the hard problems, we want to be an easy button about the data management. We’re not going to solve every machine learning problem in the world, but we want to plug into other platforms f to let people invent and be bespoke about specific parts. Our goal is not to take over the entire machine learning world, it’s really to solve a problem that a lot of people have to do themselves, and then build a platform for people to integrate with.

Additional topics discussed:

• Marple’s background in software development 
• The Market for Unstruk
• Insights on being a customer facing platform
• Managing IP and patents

Host / Podcast Bio: Eric Wright is a Systems Architect and host of “DiscoPosse”, named from two words that hold special meaning to Wright’s history and personal interests. DiscoPossee is a show meant to simply be about life and technology, the latter being Wright’s “primary love.” 

Guest Bio: Kirk Marple is the CEO of Unstruk. He began his career in Media Management and entertainment, owning a transcoding company. After selling it to diversify his prospects, Marple found inspiration within other areas of data management such as the automotive industry. To Marple, the use of data for car imagery and closed captioning had parallels in “time based telemetry.” However, Marple noticed problems within the storage of unstructured data and metadata within file based formatting and decided to expand into the business of optimizing a system to store, organize, and analyze metadata within pools of stored information.

Resources and links mentioned in the show:

• https://twitter.com/kirkmarple?lang=en
• https://www.linkedin.com/in/kirkmarple/
• https://www.unstruk.com/
•