Back in the Lab: Building New Software Business in Fidelity Labs

Maureen Olejarz Welcome to Tech on Deck podcast brought to you by Fidelity Investments. I'm your host. Maureen Olejarz Domain Leader software Engineering and. Adam Ely Adam Ely, Chief Information Security Officer. Adam Ely Each episode takes listeners inside the walls of a fintech industry. Maureen Olejarz Anything from cybersecurity, artificial intelligence, cloud and crypto to the intersection of product and technology. Adam Ely Tech on Deck breaks down the topics top of mind for technologists today. Maureen Olejarz Plus, we'll give you insight into the exciting and challenging careers in Fintech. Maureen Olejarz Thank you so much for tuning in to this episode of Tech on Deck. I'm Maureen Olejarz, head of software engineering at Fidelity. I wanted to share a special message with you today. Maureen Olejarz Before we get to our special guest for this episode, Maureen Olejarz Adam Ely, who's typically my fun co-host, is unable to join us for today's session and we’ll be catching up with him next time. Maureen Olejarz In the meantime, Maureen Olejarz let's move forward. Maureen Olejarz I'm thrilled to be catching up with Vall Herard, Maureen Olejarz and Vall has such an interesting and unique career journey and is incredibly passionate about the work he's doing here at Fidelity. What I'd like to do is share a little bit of Val's background and then we'll get into speaking with him. Maureen Olejarz So, Val is the managing director of what we call Fidelity Labs, our in-house tech incubator that is building new software businesses and building leading capabilities for the industry. Maureen Olejarz Our listeners might remember an episode from season one with Mona Vernon, who was the head of our Fidelity Labs at that time. Maureen Olejarz In the spirit of mobility with Fidelity. Mona has taken on a different role here, but we just wanted to connect this episode back with Maureen Olejarz the time that we first got introduced to Fidelity Labs back in season one. Maureen Olejarz So, Val, welcome to this episode and to the show overall, we're excited to share your journey with our listeners. Vall Herard Thank you for having me, Maureen. I'm very happy to be here. Maureen Olejarz All right. So Val would be great if you could tell us a little bit about Maureen Olejarz maybe your journey in Fidelity Labs and how that all started. So we just referred back to one of our episodes from season one, and Mona Vernon was one of our special guests as head of Fidelity Labs at the time. And the both of you met. Maureen Olejarz And it's part of your story of beginning your career journey here at Fidelity. I'd like if you could share a little bit with us. Vall Herard Yes. So, as a matter of fact, Mona and I, we worked closely together for years. In fact, back in 2019, she was the one who told me about the opportunity at Fidelity. We actually met at a conference, and she was explaining to me some of the stuff that she was working on as an innovator at Fidelity. Vall Herard that really caught my attention. Vall Herard consequently, Vall Herard Mona and I talked about the possibility of me joining Vall Herard Fidelity Vall Herard and more specifically, Fidelity Labs. I think some of the data science and technologies that they were looking at at the time was really attractive to me. And I decided to make the decision to join Fidelity Labs because of Mona. Maureen Olejarz Excellent. So Maureen Olejarz we all know Mona and Mona can be quite persuasive. I'm glad she persisted in pursuing you for a career here at Fidelity Maureen Olejarz as managing director of Fidelity Labs. It would be great if you could just Maureen Olejarz give us a little more background on what your job entails and how you think about that. Vall Herard as you mentioned before, Labs is really the heart of Vall Herard a lot of the advanced technologies that we worked on at Fidelity. More specifically, our mandate is really to build new Vall Herard technology, businesses that make makes the lives of our associates and our clients better. Vall Herard And so it's really a place where if you're looking to work on technology, on advanced technologies, if you're an elevator, if you are looking to build new things, the lab is an incredible place to join. Vall Herard So myself into a team, for example, right now we are focusing on a lot of AI work and more specifically work related to large language models. Vall Herard And a big part of what we do is really see how we can access a lot of the data that resides at Fidelity. I think in the course of doing any Vall Herard data science work. Vall Herard The big piece is you typically spend about 80 or 90% of your time working with data because data is really the new currency. Vall Herard There's a lot of talk about Vall Herard AI being Vall Herard really the electricity that's going to power a lot of businesses moving forward. But I would flip that around and say it's really about data and more specifically, access to high quality and accurate data. Vall Herard And this is one of the areas where access to that sort of data from a well-diversified financial services company such as Fidelity can really make a difference in terms of the type of technology in the type of applications that you are able to build. Maureen Olejarz So, Val, as we've been talking about Maureen Olejarz the product that you've been working on, the areas of opportunities and things of that nature within Fidelity Labs, Maureen Olejarz I know you also are passionate about what's happening in the broader Fidelity Labs. Can you share, you know, a few other spaces with the audience? Vall Herard Yes. No. So. Thanks. Thanks for that. I mean, I think what we're all working on is a small part of a larger portfolio within Fidelity Labs. So there are a number of startups that we have within the labs. Some of them are at market. Vall Herard So we have a startup called Catch Light that's intended to help advisors really find potential clients to help them grow their businesses. We have several more that are going to go to market in the coming months. And so we've built a digital transfer agent called APRA that's about to go to market as well as other capabilities that are on the drawing board. Vall Herard And so Vall Herard it's really more of a portfolio approach as opposed to individual start ups. Vall Herard within labs, Vall Herard we are always thinking about capabilities, as I mentioned before, that go beyond just what you might be working on right now in terms of really working across themes in a very collaborative way to support one another across the portfolio. Vall Herard So it's small portfolio approach as opposed to individual startup approach. If you will. Maureen Olejarz Got it, Got it. That's great. Maureen Olejarz Yeah, So well, that's really helpful. And in a little bit, we'll pivot and start to talk about the career in tech journeys that are possible. in Fidelity labs. But I'd love to just pull on the thread here that you were talking about a great way and thinking about the value of the depth of data that Fidelity has and how it can contribute to how we use ai. Maureen Olejarz Would you mind just telling us a little bit about, you know, what surprised you most as you came here and started seeing the depth of that data and and in working here? Vall Herard one of the things that was very fascinating to me when I came to Fidelity is really the breadth and depth of the data and the care that was that the firm took in really storing the data across a number of the business units that we have and the way in which Vall Herard privacy and data governance and data security were all taking into account. Vall Herard So whether you are speaking about the personal investing business or the workplace investing business or Fidelity, institutional or Fidelity Charitable or Fidelity International, for that fact, Vall Herard the abundance of data and the way in which the governance structure around that data was set up made it easier for us to build some of the capabilities that we're thinking about building. Vall Herard And so to my surprise, having worked previously at fairly large institutions where Vall Herard although they are big and diversified, they really didn't have this. The sort of data and the data wasn't stored across the number of decades that we have to sit at Fidelity. And so when I first came in and we started thinking about Vall Herard some of the capabilities we were looking to build, it was very important for us to really do an assessment in terms of, okay, we have this great idea. Vall Herard Where are we going to get high quality data to be able to build something. And so consequently, we started looking at some of the internal systems within the firm. Vall Herard And Vall Herard to my surprise, I would say that Fidelity is one of the companies Vall Herard that's done Vall Herard a really good job, if you will, of preserving data Vall Herard in a way in which it's accessible so that you can build advanced technologies. Vall Herard I've worked in the industry as a consultant. I've worked at a couple of fairly large banks on Wall Street prior to coming to Fidelity. And what surprised me the most is Vall Herard really Vall Herard the quality with which the firm and the care that the firm had taken in really building the sort of data repositories that would allow me to do data science, because as I mentioned before, about 80 to 90% of the time over data sciences is around taking the data, claiming it, making sure that there's data governance and data security around the data. Vall Herard So being able to leverage a lot of the existing infrastructure that Fidelity had built over the decades around data really made the job easier for myself and the team. Maureen Olejarz I think that's great. And in there you talked about Maureen Olejarz the work that's been done and how important data governance, cleaning the data, things of that nature. You also are discussing ideas about how to bring innovation to the market. Maureen Olejarz You've talked about AI, and I'm assuming both of those elements were a factor in making Maureen Olejarz one of the products out of Fidelity Labs Maureen Olejarz that you've worked on, which is called Saifr. Maureen Olejarz Could you tell us a little bit more about that and what it's all about? Vall Herard So with respect to Saifr for as you very well know Vall Herard in financial services, anything that's an investor and that could be a retail investor or that could be an institution, also Vall Herard anything that the investor can either hear or see or read from a financial services company is subject to some communication compliance rules. And the reason why these rules are in place is to ensure that investors are protected. Vall Herard But the process of reviewing Vall Herard anything from a sales brochure to a sales presentation, to even a speech that a senior executive from the financial services company is going to be giving at a conference is subject to compliance review, and that process can be time consuming. Essentially, you will have someone that's in marketing create the content and afterwards that gets submitted to a compliance team for review. Vall Herard And there's a lot of back and forth. So the idea behind Saifr was really fairly simple. It's like, could we create a spell check for compliance so that people who don't have the compliance background, who are creating that content, they can have a way in which that in which they can get suggestions in terms of how to set things in a compliant way. Maureen Olejarz Yeah, Vall that's really helpful. You know, as I think about my own roles and working in technology with our marketing groups, I can see the efficiencies that you would get. You know, we had the whole EY review process and everything and the more you can speed that up, automate it, be able to find patterns and things of that nature I think would be really helpful. Maureen Olejarz So for all the companies out there that are in a heavily regulated business would agree that saving time, you know, is of benefit to them. Maureen Olejarz Can you take us back to the beginning? You mentioned before about Maureen Olejarz trying to come up with some ideas of things that you could bring to market. And just talk to us a little bit about that time frame. Vall Herard I started at Fidelity roughly Vall Herard around the end of Vall Herard first quarter of 2020. So really right before the pandemic started. Vall Herard As a matter of fact, I had been Vall Herard at Fidelity for about three days before we all started to shelter in place and working from home. Vall Herard And so it was very hard to get to know my colleagues being new to the firm around that time. Vall Herard So as we were thinking about Saifr and we're thinking about Vall Herard what are some of the ideas, what are some of the things that we could build? Vall Herard One of the considerations was the fact that if we're going to build this it, we're going to have to hire a team Vall Herard over Zoom. Vall Herard And the whole project needed to come together over Zoom. Vall Herard The fact that the company Fidelity has in place the technology infrastructure that made that happen Vall Herard I talked earlier in terms of some of the data infrastructure. I would say that some of the technology infrastructure that exists at the firm to allow you to build things fairly quickly in a startup mode when everyone on the Saifr a team, well essentially hired over Zoom. Vall Herard I had not met my CTO for the first year of the company. As a matter of fact. And so that's all testament to the technology infrastructure that we have in the firm that made all of that possible. Maureen Olejarz And so that time during the pandemic and having you start here, you know, was just a different a different place in time, I take it. So as you've all moved forward and released release capabilities to the marketplace, Maureen Olejarz are you seeing impact in the industry? Are you seeing further interest? How's that all going? Vall Herard Know, I would say that the response from the industry has been tremendous. I think that we've been at market since January 2022, so not even that market for two years. And the response has been tremendous, both in terms of client client acquisition as well as under all that we've been able to capture within, let's call it that 22-month period. Vall Herard But moreover, I think one of the more interesting use cases that's come about is the fact that Vall Herard very early on we were thinking about large language models. But it really wasn't until really over the past year or so, once open AI released ChatGPT that I really started to see that the general public became aware of the capabilities around these large language models. Vall Herard But one of the more interesting elements is the fact that it's very hard to get these large language models to produce compliant content or content that is consistent with the rules that you have in financial services. And so what we've started doing and what some of our clients are doing is really taking Saifr and wrapping that around their large language models in a way in which the output of these models are compliant. Vall Herard And so whether you are using it as a way to generate content, that's it. To write a blog or to write an article that needs to be compliant by wrapping Saifr around that large language model, the output can be made compliant, whereas if you take the model out of the box by itself, it's very difficult. And we've tried, it's very, very difficult to make those models compliant. Vall Herard Whether or not you're using it as a chat bot that's going to be answering customer questions, for example, there's the same application, the need for you to make the output be compliant. And so the application of Saifr in those use cases is one of the more fascinating aspects of a lot of volume implementations over the past year. Maureen Olejarz Really helpful and so as I think then about the technology itself. Right. So part of a, you know, what our audience really enjoys hearing about from our guests is what are technologists, You know, what are you thinking about when you hired technologists and what are your technologists able to work on? Think about problems to solve and maybe just share a little bit there. Vall Herard because of communication from Vall Herard companies and in regulated industries, come in all shapes and sizes. So whether it's written content, whether it's videos, whether it's, you know, an image, for example, we need to build capabilities that's able to operate across all of those modalities because someone, for example, might Vall Herard use an influencer and post a video on social media while the company still has an obligation to make sure that this is compliant. Vall Herard And the process of reviewing that video, for example, Vall Herard has a lot of elements to it, right? What sort of imagery is in the background of the video that could be non-compliant or it could be suggestive in some ways? Vall Herard And then the translation of that video in a financial services context. So, for example, on financial services, use a lot of jargon that's very specific to the industry that a general translation capability may not lead to the most accurate translation. Vall Herard And if you are going to run a model that's looking at the context, right, Vall Herard you need to make sure that that translation is accurate from a financial services standpoint. And so we've built models that will do translation of audio and video taking into consideration the fact that there are very specific terms in financial services that might have different meanings. Vall Herard Right. Vall Herard Another element of what we were working on is the fact that we work with international companies that operate across different Vall Herard countries. And so the need for us to be able to not only correct the translation, but also to do translation into other languages is a key element of what my team works on as well. And as I mentioned to images. Vall Herard And so we are able to look at an image and really provide an indication whether or not this is something that would meet with financial services, regulation, regulatory rules. You know, a very simple case of that might be someone might have a Vall Herard drawing or a picture of a money tree. Right. Well, any compliance officer would look at that and say, well, that is very suggestive. Vall Herard Right. And so to be able to build image detection models that's able to look inside of a video and see if there's an image in the background that may be violated is a very, very awesome compliance rule, violative of some compliance, which is a very interesting problem and challenge. And so those are some of the things that the team is working on. Maureen Olejarz so this is all great. And we could do a whole series, as I'm sure you already know, on the different types of roles. You know, is it predominantly data scientists or, you know, do you categorize the the positions differently, whether it's imagery or anything else? Vall Herard Yeah. So we have a fairly diverse team given what we do. I mean, I joke all the time that we have enough compliance people in an organization to actually start a compliance department. And one of the reasons for that is that, Vall Herard yes, when you work in AI, you can look at the accuracy of a model. You can look at the confusion matrix to understand, okay, is this model accurate or not? Vall Herard But for us, it's really a measure of how well we are doing is really to compare what a compliance subject matter expert would say versus what the model is saying. And so consequently, we don't put any of our models out until we've done that human in the loop testing where we provide a set of data. So a human expert Vall Herard who's in compliance will review it without telling them what the model has come up with. Vall Herard And then we sort of like provide that data to the model, and then we do a comparison, a blind study, if you will, for the comparison to circuit. Are we in fact catching the things that an expert would catch? And so consequently, our team is very diverse. So we have data scientists that focus on large language models. We have data scientists that focus on image detection models who are data scientists that focus on sort of like videos and how you actually process tools. Vall Herard And then we have a technology organization, engineering organization, because everything that we build needs to scale. And then we have really the compliance function. And then as is typical with most technology companies, we have sales and marketing and those other functions as well. But from a technology standpoint, it's primarily data scientists who work with our engineering team. And then we essentially have the compliance team that sort of like sits over to have to make sure that anything that comes out is ready for use by the industry. Maureen Olejarz I definitely could talk all day about this, and I'm sure you could too, because this has been really interesting. What we'd like to do is Maureen Olejarz pivot a little bit and talk about your journey into technology, because as we learned about your background, Vall, there are so many interesting aspects to it. And I think as as our audience, you know, looks at their own careers and, Maureen Olejarz interests aren't linear. Maureen Olejarz I think it's really helpful when people get to hear of different backgrounds and and the ways in which people have come in to technology positions in firms. So I'd love to hear something from you here. Vall Herard Sure. So, I mean, I like to say that I came into technology completely by accident. I actually was studying mechanical engineering and mathematical economics of back then what used to be called econometrics, if you will. And I had this one economics professor at Syracuse who told me that, well, if you can solve a linear set of equations that there are these things on Wall Street called derivatives. Vall Herard And that's where a lot of the hiring is happening in the industry. And so I was very fascinated by that. Vall Herard My original plan was to be a mechanical engineer and actually build rocket ships. I was very interested in in space exploration. And so it so happens that Vall Herard I ended up getting an interview with the Bank of New York, and I started working essentially in banking. Vall Herard And I really enjoyed it. I really love what I was doing. Vall Herard when I was working there, Vall Herard supervisor, we had this one problem where, Vall Herard every quarter we were trying to focus what revenue for the division was going to be, and they were using an old accounting method, if you will, to do that. And so my boss at the time said, Vall, you have a background in regression analysis. Vall Herard Can you build a model to try and see if you can forecast the slightly better job at forecasting? So long and short of it is that I ended up building a regression model in the first few quarters that we ran it, where essentially that model had about a 10% error rate, which, you know, you might ask, Well, what's great about that, considering the fact that the accounting model was usually off by 30% or more. Vall Herard So I thought everyone thought that we had done it fairly. It's a good job. And so that's where I really started seeing Vall Herard really the Vall Herard interaction between technology and building systems. And also some of the things that I was interested in thinking about. And so from there I started getting more interested in technology and risk management. How do you actually build quantitative models in a scalable way? Vall Herard And then through that, I ended up Vall Herard leaving banking for a little bit and working as a consultant. And through that, we had built some IP that software company ended up buying, and they wanted to build a risk management platform out of the IP that we had come up with. And so through that process, I ended up Vall Herard working for that company. Vall Herard We built the platform and the CEO of the company at the time asked me to run sales and to run marketing. And so through that process, then I got into sales and marketing. And from there, prior to coming to Fidelity, I worked on four different startups. Vall Herard Some of them were very successful, some of them were not. That's that's the life of an entrepreneur and a technologist, I guess. Vall Herard But yes, and so ever since then, ever since I built that initial model, I've always had this fascination in terms of building things, trying to solve our business problems. Maureen Olejarz Yeah, it's such a great story and a great journey. Maureen Olejarz Thanks for sharing. And Maureen Olejarz I always think about that curiosity that people have is that you don't you can't always see the future in the kinds of things. But you know what you're good at systems thinking, analytical again, curious and I think the can take in many many places so thank you and so if we bring it back to Fidelity Maureen Olejarz Labs again, just to walk back through what Drew you here, I think as we talk with the audience, they they come to understand, you know, that we have our lines of business, right? Maureen Olejarz Whether it's personal investing, workplace investing is, you know, and other companies. And then there's these units that we have that really are innovative. ALM not that innovate you know there's innovation in other areas but driving change there and let's let's just talk about talk about that again just to reinforce that and then I'd like to move to something more fun and talk about your background a little bit more. Vall Herard yes. So as I mentioned before, so the lab is really Vall Herard an incubator. And so we come up with ideas and we go out to the market, we speak with people, we try those ideas to see how to do those ideas, make sense to them, not make sense. Vall Herard And so a lot of the initial work that we do is really thinking through a set of problems, either Fidelity is trying to address so that we know that the industry as a whole is trying to address and then trying to think through how can we put a solution together as quickly as possible that addresses that need. And so consequently, the teams within labs that tend to be fairly small. Vall Herard So similar to, you know, four or five people getting together and creating in your company, let's say, for example. And so we do a lot of design thinking initially around, okay, what does the solution Vall Herard should look like? Vall Herard What problem are we really trying to solve? Is this really something that can evolve into a product or is this a feature that we are trying to build? Vall Herard So we ask a lot of questions around trying to build a handle on the problem before we actually write a single line of code. And then at that point, when we think that we have something and we've done enough research, then we engage with really a designer to say, okay, now what should this thing look like? And then we do a lot of AB type testing around that. Vall Herard So it's really this idea of studying with very small teams who get very excited about the problem that they are trying to solve and who work on trying to get an MVP one out within six months or definitely less than a year so that we can validate whether or not the original thesis makes any sense. Maureen Olejarz So Maureen Olejarz behind what you're saying, you know, hearing all the things through the thread of this interview around curiosity, that system's thinking, Maureen Olejarz innovation Maureen Olejarz doesn't have to be big thinking about problems and opportunities potentially to solve. And it's it's really been helpful to hear about this. So so thank you. Thank you for that. Maureen Olejarz What I'd love to do whenever we do these podcast well is Maureen Olejarz really get to know the people that we're talking to. Maureen Olejarz So what we'd like to Maureen Olejarz in your case is talk about we understand your childhood was unique and so Maureen Olejarz it would be great if you'd be willing to share a couple of things about that. Vall Herard yeah. Vall Herard And so one of the things that, you know, that comes up about me quite a bit is that people usually have a hard time placing my accent. And that's primarily because of the fact that I grew up all over the place. I sort of speak all over the world, if you will. And so from that perspective, I think that Vall Herard my childhood was somewhat unique in terms of, you know, being in the Caribbean, being in Europe, being in Africa, in other places. Vall Herard And so through that process, I got to learn a lot about different cultures, about different people, how you know, how you comport yourself differently based on where you are, but still staying true to some core beliefs and ideas. And so through that process, I ended up going to high school in Florida ultimately, and then from Florida Vall Herard to Syracuse and from Syracuse to NYU, Vall Herard where I got my master's. Vall Herard And then from there I started working on Wall Street. Maureen Olejarz Wow. So, so thank you for sharing. It's it's really it's great. It makes you human. You are human. And getting getting to hear about that is is really nice. Was there any favorite place, Maureen Olejarz any favorite place in your background where you lived? Vall Herard I would say, interestingly enough, Miami Beach, Maureen Olejarz Okay. Well, you. Maureen Olejarz Know, Miami Beach is. Maureen Olejarz Great. Maureen Olejarz Yeah. Vall Herard the last 25, 30 years has been just incredible. Vall Herard And so I've seen it all unfold right before my eyes. Vall Herard I think that's that's been certainly was very exciting Maureen Olejarz So, Val, thank you so much. Maureen Olejarz So one for your time, two for sharing, you know, both your technology career as well as, you know, personal aspects of yourself bringing it all together. Maureen Olejarz I want to thank our audience, as always, for your listening in and getting to hear Maureen Olejarz about Maureen Olejarz people in Fidelity. Vall Herard Thank you for having me. Adam Ely Thanks for joining us for Tech on Deck. We hope you enjoyed the episode. If you haven't yet, please give us a five star rating and subscribe to the podcast on Apple Podcasts, Google Podcasts, or wherever you get your podcasts from. Maureen Olejarz Thank you to our listeners and recording studio and editors who make our episodes possible. To learn more about tech opportunities, head over to Tech Dot Fidelity Careers .com Adam Ely See you next time.