An Overview Of Foundry For AI By Rackspace (FAIR) And Its Architecture

Rackspace’s Foundry for AI (FAIR) initiative exemplifies its commitment to pioneering responsible and impactful AI solutions. By leveraging cutting-edge technology and a customer-centric approach, Rackspace Technology aims to empower enterprises to harness the full potential of AI while ensuring ethical implementation. Through Foundry for AI, Rackspace Technology endeavors to revolutionize industries, drive innovation, and shape the future of AI-driven solutions.

To help us understand the origins and the future of Rackspace Technology and more, we have with us Srini Koushik who is a seasoned leader, currently serving as President of Technology and Sustainability at Rackspace Technology, where he oversees technical strategy, product development, and thought leadership. With a rich background including roles as Vice President at IBM and Chief Information Officer at Magellan Health, Srini brings extensive experience in driving digital innovation and hybrid cloud strategies for global organizations. Beyond his corporate endeavors, Srini is actively involved in the tech ecosystem, serving on advisory boards for esteemed ventures. Based in Columbus, Ohio, he finds balance through world travel and pursuing his passion for music, particularly mastering the acoustic guitar.

In this interview, Srini shares profound insights into the post-pandemic work landscape and the changing dynamics of technology. Reflecting on his career journey, he discusses his accidental entry into technology and his passion for understanding how things work. He elaborates on Rackspace Technology’s founding and its innovative initiatives like Foundry for AI, emphasizing the company’s commitment to responsible AI implementation. Srini also explores the development of intelligent copilot solutions and Rackspace Technology’s unique approach to enterprise AI, underscoring the importance of ethical AI adoption and user-centered design.

AIM: What are the differences between working in the Bay Area compared to working elsewhere? Is there something specific about the Bay Area that uniquely impacts work culture or opportunities, or is this perception exaggerated?

Srini K: I believe that after the COVID-19 pandemic, your physical location of work will no longer be a determining factor because it has opened up the world. The best minds are spread out everywhere and not just confined to one place. I recently returned from India and discovered some of the best people working in my field are located there, so distance is no longer an obstacle. However, to answer your question, I think what made Silicon Valley and San Francisco unique in the world of technology was the ability to walk on the streets and bump into several people in the field. But in Ohio, where I am located, the economy is more diverse, and not every person you meet is necessarily in technology. Nevertheless, I believe that the exclusivity of San Francisco and the Bay Area in the technology sector is slowly changing, as I can now connect remotely with some of the best minds across the globe.

AIM: How did you find your path into a career in technology?

Srini K: Like most of these things, I’ve enjoyed every moment in technology, but how I ended here was more of an accident. I grew up in India, so you had to be a doctor or engineer or follow that path. So, I was going down that path, but then the thing that I always liked doing and still enjoy doing, which is why I’m doing this all these years later, is I always enjoyed finding out how things worked. If I see something going on, like now, I’m looking at the camera and going, “I wonder how that recording is being made and how the sensors are picking it up and converting it.” That “How” component, how things work, was fascinating. I’m the type of guy who will look at this stuff and then try to take it apart so that I understand how things work. Once you understand how things work, the things you can build with them can become much easier. That was the passion. My parents wanted me to be a doctor, and as I was going through that journey of being a doctor, I thought you can’t take apart things and put them back. Well, you can, but it can be very dangerous. But in the late 80s, growing up in India, computers were coming in, and by pure accident, my dad said there is this new thing called computers. I’m not sure how they work, but you should check it out. And so, a master’s degree program was being offered by the university, and that was the first year it was offered. I went in, did the entrance test, and got in, and I haven’t looked back since then. This gave me the perfect way to explore that curiosity. When I see a program working, I wonder how that works. Algorithms came naturally, and coding became second nature. I stumbled into it. But it’s been a phenomenal journey as I’ve gone through it and a learning journey for me. One of the greatest things about being in technology over the last 30 years is that it has changed yearly. Every single year, something gets better. And over the last two years, it’s gotten worse or better, whichever way you want to look at it. Because with AI, things are changing every single day. That constant fascination with how things are evolving and how you can use it has kept me to where I am, and that’s why I enjoy doing what I do.

AIM:What was the journey like in founding Rackspace Technology and spearheading initiatives like Foundry for AI? 

Srini K: I did my master’s degree back in India in 1987. I completed my thesis, which was on artificial intelligence. But at that time, it was rudimentary. The journey started then, but I think AI, over the last couple of years, just became mainstream with the introduction of large learning models and GPT and others. Everybody wants to talk about it and be in it. One of the not-so-good things about the industry we are in is that until yesterday, probably nobody talked about it. And then suddenly, everybody’s an expert. Magically, it happened where they went from being novices in it to creating Centers of Excellence and expertise, etc. To me, what was more real was that this technology was evolving every day, and the best thing you can do is stay involved in the technology and get your hands on it. So, if you think back to the Industrial Revolution, like 200 years ago, foundries were where you would bring in raw materials like steel and the machinery you used to create things of value for customers. When Rackspace Technology was moving into this space that resonated in the back of my mind. It’s all about creating things of value and foundries where you have the blue-collar, getting your hands dirty, and getting things done work. So, the concept of Foundry for AI by Rackspace Technology came in, and it was a good coincidence that F-A-I-R was FAIR because AI is all about implementing it responsibly in organizations. So, having that FAIR implementation was a happy coincidence. But the word that attracted me to what we’re doing is the Foundry. That has helped us because I tell my customers that if you’re looking for an expert, don’t talk to me because this space is evolving every day, and don’t believe anyone who says they’re also an expert in this space as well. But what I can tell you is what technology looks like today, what can be done with it now, and how do you set yourself up to grow with the technology as it goes forward. There are so many opportunities for using that, but the happy coincidence that I enjoy doing is what’s different about the way Rackspace Technology is going about it: we want to make sure that when AI is implemented, we’re doing it responsibly.

AIM: What’s the story behind the name “Rackspace,” and does your company offer hardware solutions?

Srini K: We’ve been in this business for 25-plus years. This organization has been very innovative from 25 to almost 30 years ago when the whole Internet took off, and people were starting to put their websites on the Internet and do transactions on the Web. It was tough to support these applications because web technology was pretty complex in those days. Rackspace Technology was formed to do manage hosting of these websites. It was brand new; nobody else was doing this at that point in time, and we did that. And as that has continued to evolve, when you think about it, if I’m managing your software, application, and websites, you’re buying space in a rack in a data center. So, Rackspace Technology, the origin of it, came back to the fact that you don’t have to build or manage your own data centers but just buy space on our racks and don’t worry about those mechanics. We’re the experts who will take care of it. 

That was how this company was founded. Despite advancements in cloud technology, we’ve remained true to that. We’re an engineering and technology company so technology is our core. We are people who help implement solutions for customers so they can focus on what’s important. If you’re an insurer or a bank, insurance and banking is your main business; you shouldn’t have to worry about all of the complexities of managing technology, we can do that. So, that blue-collar foundry let us handle the technology complexities is what this company was founded on. And for us, the Foundry for AI by Rackspace (FAIR) was a natural extension of what we’ve done for 25 years. 

For Rackspace Technology, the first thing we tell everyone is that we’re a customer-first company because we exist because of our customers. So, it’s the Customer First and Cloud First. Customer first because we want to do the right thing so that the customers can deliver what they are doing. And Cloud first because we are a cloud company. 

When we say Cloud first, we help our customers build applications and solutions on the Cloud. There are two specific deployments of the Cloud. We work with AWS, Azure, and Google Cloud in a public cloud setup. So, when we build the solutions, we can deploy them on their cloud infrastructure. The other option that many of our customers are asking for is a private cloud option, which they want to run this thing with cloud-like characteristics but in a single tenant mode in that piece. That’s where we provide hardware, operational support and others. On the private cloud side, we provide the entire stack of infrastructure and platform as a service and the applications that run on top of it. On the public cloud side, we build data and application solutions, and we deploy and operate them on the public Cloud, AWS, Azure and Google. So the entire tech stack with the focus on the Cloud.

AIM: Could you provide insights into your approach and how the “ideate, incubate, industrialize” framework applies to your projects?

Srini K: AI is complex as it is. If you pick up any Journal Magazine and read about all of the developments, it’s actually pretty complex. There’s no added benefit to complicating it further when you’re trying to help customers. So, we first wanted to make sure that we made AI approachable and say that it’s not complicated to use and implement AI. The product development process you have to go through is that you have to come up with an idea (Ideate), and you have to be able to prove that that idea works in your space, that is, the incubate side. 

It’s one thing to have an idea and another thing actually to convert it into something that is actual. But let’s not forget that the result is not to build a toy, pilot, or prototype. It’s to industrialize it because the only way AI becomes useful to any organization is when you start using it and getting the benefits of it. That’s the industrialized portion.  The process is self-explanatory, but you have to go through the process of ideating something, incubating it in your environment, and industrializing it because when you complete that cycle, you can start delivering results.

AIM: Could you share some examples of other cool projects you’ve worked on using the “ideate, incubate, industrialize” framework?

Srini K: There are several. One of my favorite ones was in Singapore. We work with a not-for-profit, nonprofit organization that focuses on helping people who are on the verge of committing suicide, it is a suicide hotline where you can call and talk to someone. Basically, when you’re in the worst phase of your life, when you talk to someone, you’re able to get that counsel, and you can come back from the edge, so there’s nothing more noble than trying to save human lives. This organization does that, but one of the challenges this organization and others like is that it works with those who volunteer so there’s a lot of attrition in terms of the people who are on call. So we leveraged AI and generative AI to look through previous call transcripts. When was it done the best way? What types of conversations should you have? It never replaces a human being, but if I have someone who’s brand new volunteering on the suicide hotline, you need guidance on how to have that conversation, and we have an AI assisting them.

So, it’s essentially helping human beings by ensuring that AI is used for good. The solutions don’t get any more satisfying than that. You can always make things more efficient and save money, but that was a cool project. We enjoyed working with that organization, and we can start doing more and more in that space. 

AI becomes real in organizations when delivering value because the cool factor wears out very quickly. So, we’ve shown it, but if it’s gonna cost me a lot of money and if it’s not of any use it’s not useful. So, we’ve used it to simplify complex areas. One of the challenges in the industry today is there are not enough subject matter experts, no matter where you look. It’s not enough subject matter experts in technology. Not enough subject matter experts who can help you navigate the regulatory guidelines because they keep changing and evolving. What’s happened over the last 15- 20 years is that with the advancements, the domains have become so complicated that it’s very difficult to develop enough humans to catch up with it. So, many of the solutions we have been implementing in our first wave of AI have been about how you learn from that pool of knowledge. If it’s compliance, how do you learn from that and assist that human being to make better decisions? That’s a really good use of AI because these domains are going to get more and more complex, and new things and regulations are going to come out. And it’s tough for us to stay updated with all the changes.

AIM: What do you consider to be some of the most urgent challenges facing enterprises today, particularly in the realm of AI? Where do you believe they require the most support, especially concerning Generative AI?

Srini K: One area in general is having subject matter experts who know how to implement these solutions, which is a challenge. Many people are wrestling with the impact of AI in an organization. I’ve seen analysts describe it as the next big thing, we saw Cloud, and I think they’re grossly underestimating the potential of AI because everything that you’re seeing, both the good and bad impact of AI, is real. I think the potential for AI to change who we are as human beings in a positive way is absolutely real. In the same way, all the bad things they’re projecting are also real. 

One of the big things that most organizations are struggling with today is how to adapt these technologies. How do you adapt AI in a responsible way in my organization, but also do it so that I don’t put the business at risk? Generative AI is fantastic. Every one of us has used ChatGPT. But the error rates and hallucination rates in ChatGPT are still pretty high. It’s great when you’re writing an article, or at least writing the first draft of an article that you will review later and then edit, change, and improve it. That’s a fantastic use case. 

But if you’re going to use generative AI to answer important questions, where the high stakes are there, then the hallucinations and the error rates you have are too high to be accepted. So, it doesn’t mean you stay away from technology, but how you adopt it in a responsible way to help human beings is something that most organizations are wrestling with. They want to use it, but they’re trying to find the best way. The way we’re trying to help them is you must think about AI as an employee in your workforce. If I’m putting AI into Rackspace Technology, AI is just like any other employee I hire. If I hire an employee that doesn’t match my company’s values, and the employee doesn’t follow rules and values then bad things can happen. Companies are used to dealing with that. They’ve got to adopt the same attitude with AI because if I’m going to bring AI into it, how do I train the AI to operate with the values that we have as a company, and how do you govern it to make sure how do you put the guardrails and governance in place so that it doesn’t veer away from what we wanted to do. 

AIM: If you had the opportunity to ask VPs of data science at leading consumer organizations one question regarding Gen AI and its implications for improving sales strategies, what would that question be, and what insights do you hope to gain from it?

Srini K: It’s always going to be, what is the purpose that you want to use AI for? The core of what you need to implement starts there because today, technology has become so powerful that you can create anything. But I think just like using an iPhone to record things and others, Steve Jobs and others pioneered the whole thing and it’s nothing new, but there’s something called user-centered design. In the user-centered design, there are three components. If you’re going to build a solution, is it desirable? Do people want something like that? The second thing is, if it’s desirable, is it feasible? Someone may want something, but can I build it? Then the third component is whether the technology is viable. You may like it, and you may want it; I can build it, but can you make money off of it, or can you have the desired outcome? And that simple thing hasn’t changed.

So, it is desirable for sure. Things are desirable because nowadays, everyone likes the cool factor; they want to say they’re working in AI and Generative AI. Desirable is not an issue in AI. Actually, increasingly feasible is not a not a problem either. Because the rate of technology is changing every day, there’s something new coming. If it wasn’t there yesterday, it’ll be there in two days. It’s that fast. 

When desirability and feasibility are no longer the issue, viability is most important. What is it that you are building, and why are you building it? It is more important. What happens is that it is forcing a lot of technologists and data-focused people to step away from their comfort zone and force them to say, why and why am I actually building something like this, and what’s the purpose of that? And I think the more people we can get into that discussion at this stage, we’re going to have way more successful implementations of AI, and we’re gonna have AI that’s responsible.

When people are happy with what they use, they take it further and say it needs to be desirable. That’s the idea behind that user-centered design framework, which has been time-tested, and it’s been a good one. When we went down the whole Foundry path, we said that we don’t have to reinvent the wheel. We don’t have to create new terms, and other types of things that already exist, but how you apply that has to change. Because the technology and the desirability of this technology have becomes very high, it’s through the roof. Everybody wants to be in AI and everybody wants to build something.

AIM: What’s next for Rackspace Technology after developing the intelligent copilot for enterprise? How does this achievement fit into your strategy, and why formalize it into a product? Is it an accelerator for future endeavors?

Srini K: The intelligent co-worker for the enterprise ICE that we built is to solve many of the problems we’ve been discussing. We have 25-plus years in this business, and so much information has been gathered over these 25 years. It’s stored in repositories, PowerPoints, articles, and wikis. If you’re a new employee joining us, it is very difficult to find out where all of the sources of information are but it’s not just that. For someone like myself, I’ve been in this company for two years. I don’t know everything that exists in this company. The sources are vast. So, the intelligent co-worker for enterprise is we’ve invested a lot of this knowledge into our solution. We built it on top of a large language model and put a natural language interface in front of it. So I can come in and say, “What solutions do I have in VMware?” And it goes and picks all the latest stuff. It’ll tell you, here’s what we can offer our customers, and here’s the source from which I got that information. So, it is a companion to anyone coming in to be able to use that. And as the knowledge base for Rackspace Technology continues to grow, that will be a companion. And for us, it was interesting to build that because people must treat AI as an employee. So we had to train the model so that it behaved in and responded in a way that’s consistent with our values as a company of a customer first and cloud-first. How do you build that into the training to do that? So, once we built it, it started to serve three different purposes. One, we can use it internally. Number two, back to this foundry aspect, is that I’m not talking to a customer from a PowerPoint; I’m showing them actual code. So that starts to build credibility and confidence in our customers that we can do it. Then the third component is that if we did the first two, we could make it an accelerator. So, we’ll extract the key portions of our design and make it available. If you go to the AWS marketplace or Azure marketplace, you can see the ICE accelerator sitting there. It’s all three of those. 

AIM:What sets Rackspace Technology apart from others in the development of copilot solutions?

Srini K: Everybody’s building co-pilots, but the people who are building co-pilots today are building them to enhance their existing software stack. So, if I’m Salesforce, I’m building Einstein to make everything in Salesforce available. If I’m ServiceNow, I’m building the same thing for myself. You’re going to see a variety of co-pilots pop up in the market, but we’re coming at it from an enterprise standpoint. 

Rackspace Technology uses Salesforce and ServiceNow, and they’re all getting their co-pilots to us. The angle that we’re coming in with for Rackspace Technology, i.e., for any of our customers, you have to think about these as employees coming into your environment. How do you get them to collaborate? So, it’s almost like when you’re building a team, and you went out and hired people from six different companies to come in, and everyone’s coming in, say, hey, look, I came from IBM, and at IBM, we used to do it this way. Then, the next guy came in for some other company, and while all those perspectives are valid, they’re irrelevant to your enterprise because that was in the company, they were in. So, what we’re trying to do and what ICE is different with the accelerators is you’re trying to build that for that enterprise and integrate these other things in so that the experience is seamless and that the AI that you’re putting in is an employee of your company, or at least behaves like an employee of your company.