Bill Kerr: IT jobs have always required at least a passing familiarity with the business’s mission and market. But with the maturing of cloud computing, more firms are looking to their engineers and developers for hybrid skills: mastery of both technology and business. This is especially true in the financial sector, as Wall Street firms see opportunity in parlaying their own cloud migration into products and services. What are the implications for business strategy, the job market, and training?

Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Bill Kerr. I’m joined today by Marco Argenti, Chief Information Officer of Goldman Sachs. Marco joined Goldman in 2019 after overseeing cloud services at Amazon Web Services. We’ll talk about the changing role of the developer, from a focus on systems and infrastructure to solving business problems, and how this impacts the C-suite. We’ll also talk about what it takes to attract and retain top talent and the emerging hybrid skill set. Welcome to the podcast, Marco.

Marco Argenti: Thank you for having me.

Kerr: Marco, let’s speak with a little bit of your background and what attracted you to the CIO role at Goldman?

Argenti: Throughout my career, I’ve been always in technology. In my tenure at AWS, which lasted for about six and a half years, I worked and I launched several areas, such as mobile, serverless [computing], Internet of Things, messaging, augmented and virtual reality. And that gave me an appreciation for the concept of technology as an enabler for digital transformation for companies. Especially in the Internet of Things, I had the opportunity to really work with large industrial groups, such as Volkswagen Group, for example, to look at their process with the lens of technology and looking at really a front-end or a front-to-back transformation, where you really use technology to digitize the business, and also to really understand your business better, and really be able to apply a different level of agility to your business. So the attractiveness of Goldman Sachs kind of stemmed from my desire to go from being the provider of tools for others to transform to be actually driving the transformation from the inside on a company that is more than 150 years old and, obviously, has gone through several cycles of transformation. And so for me, the attractiveness of moving from the IT industry, if you wish, to an actual other industry, in this case financial services, and really driving the transformation from the inside was kind of a natural step after having been on this other side of the fence for many years.

Kerr: Yeah, it’s fascinating, because we think of Amazon and other digitally native companies as being built around the technology. And here, you’re bringing the technology into an organization that’s built differently more than 100 years ago and trying to both apply some of the technology into the organization, but also transform the organization to now sit more around that technology journey.

Argenti: Yeah, absolutely.

Kerr: So you’ve talked about updating the role of engineers at Goldman, from being people that maintain the infrastructure to being more problem solvers. I’d love for you to tell us a little bit about setting that priority and how you’ve been measuring the results from it.

Argenti: Yeah, so the overall observation here is that the role of engineers has really changed quite dramatically over the past few years, and definitely in financial services. The role of software is, actually, in some cases, the business, itself—is becoming the business, itself. And financial services, of course, being a fully digital business, it’s really putting technology historically at the center, but more like an enabling or a supporting back-office function that has transformed in the last few years as a driver of our competitiveness. A lot of that is about the ability to react really fast, to take a data-driven approach to decisions, to have decisions that are driven by looking at really, really subtle and weak signals in exploding amounts of data. It is the move from batch to real-time—for example, being able to settle transactions instantly instead of after days—is the incredible sophistication of some of the trading algorithms. It’s the way you need to approach “know-your-customers,” or KYC, or the work that you do to comply with a constantly changing and constantly evolving regulatory landscape. And those are really, really hard problems to solve. And so, if you look at the agenda of a CEO, a typical CEO of a financial institution, but also in other industries, technology is now right at the center of the strategic agenda of a CEO. And, of course, you probably heard and people have heard about our own CEO, David Solomon, being a very passionate advocate for technology being at the forefront of our strategic agenda. That, of course, requires an upskilling of developers who are, first and foremost… developers need to really be embedded in the revenue-producing businesses and divisions. And the other one is that it requires developers to really understand the customer and the business and the why—and not only the how—about ROI or about customer experience, about metrics of retention and adoption, and metrics of effectiveness of algorithms, and our ability to be able to effectively serve our customers at the end of the day. And so that has created a whole new discipline on how you approach software development, which is fundamentally based on the question of who’s your customer? On the question of, what are the problems that you’re going to try to solve or the opportunities they are trying to capture? And how do you know those are the problems or the opportunities? And how do you go about the customer experience? And really capturing that in something that we took from the Amazon playbook, called “working backwards from the customer,” i.e., empowering developers to really understand the business goals from the beginning of the process, before they write the first line of code. And so that really changes the role of engineers, into more like someone that uses logic, uses their ability to process very complex information and very complex framework and mental models, into a business context. And then helping other decision makers to navigate their way through the complexity. It’s almost like the reinvention of software management science in a way that is centered about pivoting the approach that you go about when solving problems—taking customers, putting them first in the hierarchy, and then working backwards from their need or from their opportunity into then the specific solutions to a problem.

Kerr: That’s so thoughtful. Let me ask almost like a mini-quantification question here. And I want to put—let’s say at zero—where financial services or automotive or all the kind of industries that you’ve traditionally worked with before they really went deeply into this software future. And then on the other end, you have the tech companies that you worked in as part of your career. How far in that left-to-right, zero-to-10 spectrum do you anticipate financial services sectors to go? Is it going to be somewhere split in the middle? Is it going to end up leaning even more toward the tech companies than not, to where they began? How would you quantify that?

Argenti: I think the endpoint is probably quite far, meaning that really the world is changing, really, in front of us at a pretty rapid speed. It’s ultimately your ability to really reason in real time with large quantities of data. So I think all the industries are going to arrive to a point in which they’re going to have to shift to the left quite a bit with regards to the sophistication of their IT department, compared to what it used to be historically. By the way, there is a little bit of a parallel there, because some companies that really get that are going to become providers of others, even within the same sector, and it’s already happening today. The idea is that not everybody’s going to be able to develop their skills. So this spectrum that you mentioned is actually going to be different, depending on which players you pick within the industry. Some companies or some financial institutions are going to be more toward the traditional approach. Some others are going to lead. And those who are going to lead will become providers of technology to others. And that’s also one of the fundamental changes, because it actually creates a business opportunity, which is almost like the next evolution of the concept of cloud, where you are ultimately developing skills on managing complexity and on abstracting complexity that become a product, because it’s valuable to others.

Kerr: I think this is a fascinating change that’s happening at Goldman that you are leading. And so maybe walk our listeners through how the offerings of Goldman and the way it works with clients can change as a response to the public cloud and your role in there. And then a little bit then how that translates into different engineering skills that are required.

Argenti: We recently announced the creation of a new unit, called “Platform Solutions,” which for the first time I guess in the history of Goldman, but probably in the history of many banks, puts developers, software developers, as our clients. And I’m talking about the developers that are actually working in other financial institutions or other corporations, which all of a sudden become the primary customer for their business. The reason is because, like I said before, there are areas of our business that are actually extremely complex to manage. And where that complexity can be abstracted into a product, they need to find the product experience, developer experience, as good as it can be. The first wave of this abstraction of complexity happened in the IT world. I think the same thing is happening in the other industries and in our industry with the concept of the industrial cloud, where you abstract the complexity of the business and then you offer it as service. It could be routing algorithms for the transportation industry. It could be DNA sequencing. Or it could be how you run a clinical trial for healthcare. In our case, it’s primitives in the business, such as payment, such as deposit accounts, such as checking accounts, such as issuing your credit cards, such as loans. Those are all primitives of banking that, at scale—especially if you operate internationally, if you operate across the globe, if you operate in a very demanding regulatory regime—are very hard to manage at scale. And so those become the components of those platform solutions—or what we call the “financial cloud”—and those are actually something that you can offer to clients. And so we already have some great examples of that in Goldman Sachs—the first one being “Marquee,” which is our digital storefront for institutional client services, where we deliver our unique market insights, our analytical tools, our execution services, our data services through an integrated digital platform. So imagine Goldman data mixed with your data with sophisticated analytics model, and then become a decision tool. Think about transaction banking, which is a very modern, state-of-the-art suite of products that helps corporations manage their treasury. Imagine a checking account for corporations done at a global scale, done with the ability to create virtual accounts, originating payments, tracking balances, et cetera, et cetera. Think about credit cards as a service—the fact that we partner with Apple and GM, leveraging our technology stack to really create really, really seamless integrated credit card experiences, where you really have the benefit of enjoying a complete integration in the experience—for example of the iPhone—which was not possible before if we didn’t take this developer-centric approach. And there are many others of those. They become kind of the connecting tissue of a digital economy, where a lot of the transactions are happening computer-to-computer, rather than computer-to-human. And I think that’s really where the next frontier of the business is going to be.

Kerr: It sounds like you have now even doubled down on the increased urgency to have a great cadre of developers with more-advanced skills. One was Goldman’s traditional products, and those have to be more tightly integrated with the technology going forward. Here, Marco, it sounds like there are, even front and center, the developers speaking to other developers in a B2B type model and making sure that’s successful. So how do you staff that? How much of that has been built up internally? How much comes from new hires from outside? What’s the skill base that you’re looking for?

Argenti: Yeah, so it’s a mix because, of course, in Goldman, one of the first things that I noticed when I joined was the incredible quality of the developer community that we have within Goldman. But then some of these concepts of digital transformation and some of the specific skills with regards to cloud, for example, et cetera, are something that had to be brought in externally. And, obviously, there isn’t really any formal curricula that actually forms these skills. And this is something that I’ve been discussing also with some players in academia, obviously. Those are generally learned on the job, because this is kind of a new job, it’s a new competence, a new skill of someone that is at the intersection of deep business knowledge and deep technology knowledge in a modern scalable way—mostly cloud based, and mostly mobile based, and with a big influx of data science and machine learning, this kind of perfect intersection of these three elements. And so when we recruit top talent, we really focus on these three key factors. A sense of purpose—they want to have an impact at the macro level on the financial industry because, ultimately, they participate to the transformation of the industry, itself. And also they participate directly on how the client relationship is changing. The second one is the ability to deal with extreme complexity at scale. And so I think we’re looking for people that have had the experience of tackling very complex problems that become exponentially more difficult with scale—architecturally break them into sub-components—them being modules or microservices that can operate independently, can scale independently—and a culture of trust and collaboration, and the ability to learn constantly, and to lead with data in your decision-making process are something that culturally is really the cornerstone of what we are looking for. And, of course, we have an incredible team of all the 12,000 engineers. We hired quite a bit of talent from, and we continue to do, from the external world. But really the magic happens when people coming in and bringing a fresh look at things are actually really cooperatively working with people that have decades of experience on these incredibly complex problems. And then, at that point, you are really kind of getting into this magic intersection, where the whole culture of engineering becomes a major driver of change and a major driver of strategic advantage.

Kerr: So, Marco, as we think about the skill sets, it’s not just the developers, it’s also the very senior leadership. And how does this technology inside the organization impact the most important skills in the C-suite and the team that’s surrounding it?

Argenti: I think we are at a point where the executive management of every company needs to be literate, at least on the strategic level, on what the implications of technology are in the business. Like I said, obviously those who try to approach that by saying, “Hey, let’s all take a computer science course or try to code,” that’s definitely the wrong approach. Or at least it takes a long time. The good approach is actually to have engineers be part of the decision making within the company, so that technology becomes part of the discourse at the highest possible level in the organization. And I think that’s what’s been happening at Goldman.

Kerr: Marco, our listeners will have heard the term “user experience,” UX. You created a “developer experience,” in part because this is so important to Goldman. So a developer experience team. Tell us a little bit about that and how you measure its impact.

Argenti: Developer experience not as evident to the eyes as a user interface, because generally it’s something that is a little bit deeper, but it’s equally important. And it’s equally a driver of adoption and stickiness of services. And when I say that practically, we mean the way you do documentation, the way you do your examples and your code samples, the way you do and you manage your interfaces, application programming interfaces, so that you don’t push unnecessary work to your developers. And the way you provide testing tools. The way you provide environments to evolve your software and put it into production. The way you are kind of friendly with support of different languages, et cetera, et cetera. We created a single developer portal on developer.gs.com. So that was kind of our homepage for developers where you could find everything you need in order to develop on Goldman Sachs in a single place. We formed, in 2020, a developer experience team to focus on really driving excellence around really having an informed human-centric kind of approach to how you do computer-to-computer communication, or developer-to-developer communication, the same way as you would have UX expert or user-experience experts. But I think the point is, linking that experience to also the empowerment to drive business outcomes is really a key. So giving developers also tools in order to be able to measure the impact of their work, the same way as you will provide, for example, analytics to a web developer, it’s extremely important. One of the key concepts there is that we really want to make the internal developer experience almost indistinguishable from the external developer experience. So really treating all customers with the same level of respect, no matter if they’re internal or external, because, ultimately, that drives innovation, that drives also retention of talent. And so exposing services to internal developers through also very clean, very well-designed programming interfaces, well documented. We created an internal developer hub, which we call the “Eng-Hub” or the engineering hub, where every new employee can go in and learn also through videos and through interviews and through templates, et cetera. And then lastly, so that developers don’t have to reinvent the wheel every time or do a big reset every time, to capture a lot of the intellectual property into what we call “platforms,” which are essentially highly reusable services or software that can be used across several teams. And those platforms also become the foundation of some of the services that we offer externally—for example, mobile platforms. For example, identity access management platforms. For example, incident management or observability platforms—these foundational blocks that become a way for developers to get from point A to point B in a much shorter timeframe.

Kerr: Marco, along the way, you’ve mentioned the size of your workforce, which is large, very large. Are there changes that are underway in terms of how you’re sourcing talent into the company? And we’ve spoken on several points about the development and the training of the talent. But new pools of talent that are being provided?

Argenti: Our recruiting team has done a really good job at adapting to the various changes and conditions. And, of course, we’ve done things such as opening new offices in new locations and growing them. Historically, of course, we’ve been extremely strong in places like India, where we all recently opened a new location in Hyderabad. But we also kind of looked at locations that are a little bit less obvious, like for example, Warsaw. We’ve been expanding in the U.S. in places like Seattle or Dallas or Salt Lake City. And we take more of a center-of-excellence type of approach, because we believe that sites need to have their own identity and sites need to have critical mass of talent on a specific scale. So that approach kind of informs a multiyear strategy, which can be a little bit robust or at least tries to be anti-fragile from the various changes in market condition. So we’re looking at always kind of trying to stay current there, because there are moments, like when you need to, for example, build entire new areas, et cetera, where you might need a good mix of the two and sometimes a large number of multidisciplinary type of developers. And then others are, for example, you’re kind in a more competitive or consolidated market where you need the hyper-specialization in order to keep your leadership. And so an example of that, for example, is being in electronic trading, which is a very mature market, where we have tremendous expertise in order to get to the level of service quality that our clients expect. And so, I think this is kind of something that you need to plan ahead, to keep your eye on the ball with regards to how to continuously evolve and how to continuously refine our recruiting strategy.

Kerr: Marco, given the many ways in which it is now at the heart of the organization and impacting so many aspects, how do you judge, overall, its performance? As you think about the next five years, what makes for a very strong contribution? And is that different than it was 10 years ago, when you’re a bit more in the back office, and how your performance would’ve been judged at that point?

Argenti: Yeah, it is different, because we measure our success in terms of client outcomes. And that’s really, at the end of the day, really what matters. The return of investment is a return of investment; they need to materialize in a business performance. And so really we’ve been using the same OKRs, the objective key results, and the same KPIs [key performance indicators] as the business. We made the decision a few years ago to embed developers into the divisions in what we call the “front-to-back,” where developers are really embedded into the business units, and within the business unit they share the same KPIs as the rest of the organization. So I think the more modern approach to... In addition to all the hygiene metrics or the table stakes, which are, obviously, the quality of your code and the number of incidents, et cetera, the quality-of-service objectives, we started to look at the client side, the client-facing metrics, your client expectations, your service-level expectations from a client perspective, and then how that translates into improvements of business KPIs. And that has been increasingly at the center of our strategy with regards to how we manage performance and how we really align careers and career progression, et cetera, with the achievement business outcomes.

Kerr: Marco, a final question here: Given your spot at the frontier of various technology leadership roles, what are the emerging trends that you are tracking? What are the things that are really top of mind for the next five years?

Argenti: The first one I mentioned a little bit already, but this emergence of the industrial clouds, where some of the players in every industry are becoming service providers of others, where the abstraction of the complexity and the packaging of your knowledge into software becomes a product, itself. So I think that is a trend that is very powerful. And I think I’m predicting that that will become more and more central in the strategic agenda of CEOs of any financial institution. The other one is really the evolution of big data, in terms of the ability to deal with massive amounts of data in quasi real time, or in a way that, A, can feed analytics and decision making but, B, also can flow within the organization essentially in real time. Events happen at a higher and higher frequency; data becomes sometimes overwhelmingly available. And, therefore, minimizing data movements and simplifying data models, making them consistent, and being able to process the data in real time is becoming an incredibly interesting area. And then, of course, we’ve seen a lot of recent evolution on AI and ML, and especially on large foundational models, the large foundational models. And especially generative AI, like the GPT and Dali, they’ve entered the pop culture, the popular culture right now in a pretty major way. But underneath that, there are some really foundational new concepts: the concept of transformers, the concept of attention layers, the concept of AIs that are teaching themselves how to learn, which is kind of extraordinary. That has allowed us to reach a human type of capability in areas such as content creation or image creation, et cetera. But there is definitely applicability in the field of natural language processing, and document management, and try to reason around concepts, and around documents, and around structures, which is I think can be very impactful in every industry. And I think those three trends to me are the ones to watch for.

Kerr: A lot to watch for, and I’m sure it’ll have lots of implications for the workforce and its strategy. Marco, thanks so much for joining us today.

Argenti: Thank you. It’s been a pleasure to be with you. Take care.

Kerr: We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.