On August 11, 2022, C2C will host 2Gather: Chicago, the Google Cloud customer community’s first in-person event in the Chicago area. Moderated by Lilah Jones, Head of Corp Sales, Central US, Google Cloud, the event program will feature speakers Francisco Maturana, a data architect at Rockwell Automation, Vrinda Khurjekar, Senior Director of AMER Business at Searce, and Pythian CTO Paul Lewis. The panel will discuss the technical and business advantages of using AI and ML on Google Cloud. In advance of the event, we reached out to Paul Lewis, an engaged and active member of our community who joins us from our foundational platinum partner Pythian, to discuss AI and ML insights, connecting business and technical collaborators, and the value of a peer-to-peer Google Cloud community.
Pythian has received significant industry recognition for its data solutions. To what extent today does a data solution necessarily require an AI or ML component?
It is fair to say that most data solutions have a “why,” and that why is because I’m trying to create some sort of insight. Insight might be for the purpose of creating a new customer experience, or creating some insight for efficiency, or monetizing the value of a current set of offerings, and that insight requires a combination of three things: I need to find where the data is in my core systems from my third party, I need to create analytical value in a data platform, and I need to use AI and ML algorithms to source out that piece of insight which I’ll use to make a decision. So it has all three of those components. I’d argue that if you’re starting with the end, starting with the insight, all of that technology and process is required to deliver on it.
You spoke with C2C earlier this year about cloud security and the shared roles of businesses and cloud providers. When working with systems and processes that are largely automated, what cloud security considerations arise?
Cloud security requires the assumption that you are going to bring your algorithms to the data versus the data to the algorithms––a really big shift from exporting data out of a production system into your laptop, producing your algorithms in your API of choice, and then sending that algorithm back up to be both trained and tested. Now it’s about training and testing in the cloud, which has access directly to those data sets internally and externally. So that’s the big shift. Moving where you’re actually both developing your model, training your model, and creating inference or executing on that model. It is the best bet to do that in the cloud.
A big problem in healthcare, as you can imagine, is sharing information across organizations. Since data sharing is required to make complex diagnostic decisions, I need to be able to package up that information from a diagnostics perspective, share it amongst a group of people, and then that prediction can come together. Multiple practitioners can participate in the model development, multiple practitioners can provide input into the model and the training, and then infer it for the purpose of new patients coming in.
On August 11, at 2Gather: Chicago, you’ll be speaking alongside Francisco Maturana, a data architect at Rockwell Automation, and Vrinda Khurjekar, Senior Director of AMER Business at Searce. As a CTO, how does speaking alongside both technical and business professionals influence the kind of discussion you’re able to have?
My conversations tend to be balancing the difference between why and how. On the business side, what are ultimately the business goals we’re trying to achieve? It tends to boil down to something like data monetization. Now, monetization could simply mean selling your data, it could mean creating a better insight on your customers, maybe as customer segmentation, maybe it’s wrapping a non-data related product with a data-related product. Like a checking account alongside an ability to predict spending behavior changes over time. Or it might be internal, making better MNA decisions or creating some sort of efficiency in a process, or just making general business decisions better or cleaner in a sense.
So, you can take that why and say, ‘well, that why can be delivered on a variety of hows.’ A how can be as simple as a query and as complex as the entire data engineering chain. And that’s the bridge between the why and the how. Not only does the data engineer or data architect get a better appreciation for the type of business decisions I need to be able to make based on this work, but the business person gets to understand the potential difficulties of making that actually true.
Do you think that most customers come to a peer-to-peer panel discussion with a why or a how in mind?
Yes. Very rarely is it unanswered questions. Very rarely is it, ‘I know I have some nuggets of gold here, could you possibly look into my pot and see if there’s anything interesting?’ That might have been true five years ago, but people are much more well-read, definitely on the business and the technology side. There has to be a why, and if there has to be a why, there’s one too many potential hows. What’s our best bet to the how? Data engineers, data modelers, and data scientists are the go-to person to hire. In fact it’s so complex that I now need partnerships of talent, so I might now know that I need a junior, senior, or intermediate scientist, because I don’t have that background. I don’t have that expertise, so I’ve got to lean on partnerships in order to figure that out.
Is being able to find the right why for the right how what makes a community of Google Cloud customers uniquely valuable?
Exactly. It’s also sharing in our expertise. There’s this huge assumption that I just have to acquire the expertise to deliver on my particular why or how, that I just need to learn Python in twenty-one days, that I just need to get another data modeler to understand what a bill is, what a person is, what a patient is, what a checking account is, but the reality is you have to balance expertise with experience. You could hire a bunch of people or train up your existing staff, but if they’ve never done it before, that’s where you need partnerships. That’s why you need a community. That’s why you need to be able to talk to your peers. That’s why you need to have these kinds of conversations, to balance what I think I can do with what’s actually possible, or what’s been done before.
Are there any particular conversations you’re hoping to have at the event in Chicago?
Yeah, absolutely. The conversations I’m looking to have are unique or interesting whys that I think could be compelling across a variety of industries. What I find most interesting isn’t that two retail chains have the same customer segmentation problem, it’s that you can take a customer segmentation retail and apply that to manufacturing of cookies. So, something we can reuse across these industries, because in my opinion these industry solutions are going to be on the forefront of the whys. I’m going to be able to download cookie client segmentation and then augment it for my needs. I don’t have to invent it going forward.
Do you have any final thoughts to share with the Google Cloud customer community?
I’m really looking forward to this particular event. It’s rare that we get to have real peer-to-peer conversations, so I’m absolutely looking forward to it, and Google’s a nice space to do it in, so, that’s always a bonus.
Are you based in Chicago? Do you need to find a how for your why, or vice versa? Join Paul, the C2C Team, and the rest of our distinguished speakers at 2Gather: Chicago on August 11! Register here: