Latest / CXOTalk / Data and Analytics at Goldman Sachs
Transcript
- 0:00Today we are discussing data at Goldman Sachs.
- 0:04We're speaking with Nemo, Raphael the firm's head of data
- 0:08engineering and their Chief data officer.
- 0:11I've been at Bowman about 20 years, always as a software
- 0:15engineer data, engineer rat, or want as we call volunteer
- 0:20stress. And now, as you mentioned, I run
- 0:23our data team or data, engineering team and chief data
- 0:25officer. For Goldman data engineering,
- 0:27Goldman is probably a little little bit different than other
- 0:32places. I, I, sort of organize my brain
- 0:36and my aim and sort of three buckets.
- 0:39We have our platform team, which hopefully, I'll have a little
- 0:42more time to talk about and our curation, our content and
- 0:46curation team. And also as a chief data
- 0:51officer, with my chief data officer hat on sort of our
- 0:54governance and quality teams. So that's that's sort of the
- 0:57background of the work we do. Data is so obviously important
- 1:03to financial services but it would be really interesting to
- 1:08hear your perspective on the role of data how data fits into
- 1:14the world of Goldman Sachs information is the lifeblood of
- 1:18financial services and people say well what is that?
- 1:20And and I think you know, in financial services at Goldman,
- 1:24of course you know information is is actually our currency So,
- 1:30a lot of a lot of decision making a lot of helping our
- 1:34clients, a lot of innovation is all based on what information we
- 1:40see out in the world and what information we see internally.
- 1:46And and so organizing that making sure it's readily
- 1:49accessible, making sure people could get to that data quickly
- 1:55to, to do to do their work or to help their clients.
- 2:00Starts becoming sort of like a core, a core component of doing
- 2:05business, that Goldman Sachs, you mentioned, the various
- 2:08aspects of your role, how you break things up.
- 2:11Do you want to dive into that? And tell us a little bit more
- 2:13about that. The first thing I think about
- 2:16data is that it has to be a first class asset.
- 2:20Right at Goldman. People have to believe that it
- 2:23could help solve problems. Help, our clients help us
- 2:26innovate and financial services. So, so the first thing we sort
- 2:31of did is okay, how do we want to work with data Goldman and
- 2:35again, as an engineer, that always push me to sort of think
- 2:39about data as as an engineer would think about their code.
- 2:44So I the analogy, I always say, or maybe it's not even analogy,
- 2:47but the way I always talk about data is like, you got to think
- 2:51about data, just like you were thinking about your code and
- 2:53that's sort of has pushed us to build a platform team.
- 2:58That really thinks about that. Things about data engineering as
- 3:02an engineering function. He thinks about the workflows of
- 3:05data Engineers, how we could help get, leverage and scale to
- 3:10our businesses so that they could start making decisions
- 3:13better faster, cheaper easier. And so the platform team and my
- 3:17team is really laser focused on that, right?
- 3:21How do you make data engineering workflows just seamless?
- 3:24Beautiful, and all the things that, you know, Developers are
- 3:29used to right? I ve s code completion.
- 3:35Making sure that we have like reproducibility of our data.
- 3:39Like we have code think that so all of that sort of becomes
- 3:43their first class like role is that won't is to only think
- 3:48about how do we make data awesome working with data also
- 3:54and then we have a Content team. So again like I said, A lot of a
- 3:58lot of our workflows at Goldman run on data.
- 4:03So you can imagine, you know, part of our team is a real-time
- 4:07Market data team. So all the all the data
- 4:09streaming from exchanges or different, venues things like
- 4:12that in real time, super low latency.
- 4:15You know, we're talking about like millisecond latency here
- 4:19right to get to Trader screens or to get into the Algos.
- 4:23And so there's a, there's a whole team of mine which is in
- 4:26the curation business. The content business, whether
- 4:30it's real-time Market data or things.
- 4:32Like our reference data that we use a, you know, like who our
- 4:36clients or what products, can we trade things like that?
- 4:39And so that team is really, I'd say, sort of a data engineering
- 4:44bent team that uses our platforms to sort of do their
- 4:47work, make sure the data is amazing, quality wise accessible
- 4:54and ready to go. And then, the third bit is Is
- 4:58really about the framework that we do data governance and data
- 5:02quality around like so a little bit around policy and Frameworks
- 5:08but then how do we push our platform team to then build in
- 5:11those mechanisms or controls on the platform.
- 5:15So when you do your work at Goldman, you sort of get those
- 5:18for free, I say or complimentary.
- 5:20So that's sort of the, a little more about the three buckets, we
- 5:24talked about here. What I find particularly
- 5:28Interesting is unlike most businesses which are of course
- 5:32have their their business data. You're dealing with huge volumes
- 5:39of real-time information of real-time data at which there is
- 5:45enormous amounts of money and risk at stake.
- 5:51I have to imagine that that places and additional kind of
- 5:56intense bird. Ian on you, as you're thinking
- 6:00about this whole technology landscape, I would sort of flip
- 6:04it in a positive. I don't think it's actually, we
- 6:06don't think of it as a burden. We think of it as a real
- 6:09opportunity, right? That I think, is the first sort
- 6:12of mindset shift. You have to think about when
- 6:14you're in these sort of high-stakes games.
- 6:16Yeah, of course, it's high stakes.
- 6:17Of course, things could go wrong, but really, it's more
- 6:21about the opportunities to make sure that again, we're helping
- 6:24our clients, do the right thing, we're helping the economy, we're
- 6:27hoping the Financial system do the right thing.
- 6:29So we sort of take pride in that instead of a sort of Burden
- 6:34thing. And the other one thing I'd say
- 6:37is, I actually like to say you said, huge amounts of data, I
- 6:43actually like to say Goldman Sachs were in the sort of medium
- 6:46sized data but very complex data World.
- 6:50Our data super complex, the types of data are complex, the
- 6:54speed is complex, there's everything from super real.
- 6:57Time low latency all the way to, you know, end of day batch
- 7:01processing the relationships between our data as complex,
- 7:06right. The product catalog is complex,
- 7:10right? It's not just use of whatever
- 7:14its actual financial instruments that people have have made up.
- 7:18So the complexity is really the really interesting and cool
- 7:22challenge here versus I'd say the volume what makes The data
- 7:29complex, or just describing a little bit.
- 7:31Can you drill into that? The complexity again comes from
- 7:34various angles, like you mentioned the complexity comes
- 7:37from the, from the speed is one angle one dimension, right?
- 7:41But also the relationships of the data.
- 7:43So you could imagine, you know, there's data about the stocks
- 7:49and bonds that peep that our clients want to trade for one
- 7:54example, but then there's a whole infrastructure.
- 7:57Around those like I don't know how many people here are
- 8:03familiar with financial services, but there's
- 8:05derivatives on those products, right?
- 8:07So now you're not just talking about stocks and bonds, but
- 8:09you're talking about really complex.
- 8:12Now, algorithms plus complex data elements about now that
- 8:19have a relationship that have been now made up in the world on
- 8:24those stocks and bonds so layering of this data and the
- 8:27Alexa tea. Now that people have come up
- 8:30with such creative ideas to help our clients and things like that
- 8:34starts pushing the boundaries of like how this data is related
- 8:38and interconnected and curated. Hopefully, that gives a little
- 8:43bit of a sense. You are managing these very
- 8:48large volumes of real-time data. Do you want to tell us at all
- 8:55about your, your technology infrastructure?
- 8:59It's not a question. I usually ask, but it seems that
- 9:02in this case given the the equities, the derivatives,
- 9:08everything else that you had just mentioned it.
- 9:11Seems the technology infrastructure has to play a
- 9:13very important role in the structure.
- 9:16Absolutely plays a huge role. Here, we have an incredible
- 9:20infrastructure team and you could imagine again that this
- 9:24goes all the way down to sort of the hardware Hardware layers,
- 9:27like the networking stack, the computers you use the network
- 9:34cards in the computers use. All this stuff sort of actually
- 9:40matters at At this sort of latency in scale, but I got it.
- 9:45I don't want a super over index on the Real Time stuff.
- 9:47The Real Time stuff is seen as a very important, very critical
- 9:51and important part of of of our of our world, for sure.
- 9:57And the team does an amazing job all the way from the hardware
- 10:00layer up. But we also have, you know, data
- 10:05challenges just after, you know, as the real-time data.
- 10:11As in, you know, when you do a trade right, that store also
- 10:14starts being complex and so we have built sort of a data
- 10:18platform here, we call Legend and we've actually open source
- 10:23that recently in the last couple of years.
- 10:25We've built this Tech stack over the last 10 years.
- 10:27Internally, two years about two years ago, maybe three.
- 10:31Now we open sourced it it's fully on GitHub, you can check
- 10:35it out. We gave the code to a non-profit
- 10:38and open source nonprofit called Fitness, Around GitHub happy to
- 10:43talk more about about that as well.
- 10:46Why did you open source at first glance?
- 10:50It seems like an odd decision to me.
- 10:53Because one would think that a firm like Goldman Sachs would
- 10:57want to keep all of that infrastructure to yourself.
- 11:01There's a couple of reasons. One is, as we were building
- 11:03this, we talked to a lot of clients who are having exactly
- 11:08the same data challenges. We Were and, you know, we're
- 11:12very obviously, a very client-centric firm.
- 11:17So we thought look, this has helped us so much internally.
- 11:20This platform has helped us so much internally, we'd love to
- 11:24give it a chance. Help our clients to help push
- 11:27the industry forward as well. And and so you also have to
- 11:31separate a little bit about the platform itself versus the
- 11:34content that we curate and work in the platform.
- 11:37So, we haven't yet. Or maybe never will open source
- 11:42the actual content, but we could talk about that as well.
- 11:47But the platform and the work we've done to sort of
- 11:50standardize, how we work with data.
- 11:53We thought that was so powerful. That that actually giving it out
- 11:58to the community and building a community around, that could
- 12:00actually help the whole industry and our clients and and
- 12:05ourselves you are wanting to share best practices.
- 12:10Is approaches to working with data things like that with the
- 12:14community and then of course, everybody's going to have their
- 12:17own data content within within those constructs.
- 12:20That's exactly right. The interesting part is even the
- 12:23so even if we don't give out the content, right?
- 12:26We are working with industry standard bodies, to even
- 12:31describe the data and structure the data, and the linkages to
- 12:35the data. So, it's a really cool and
- 12:38Powerful technique. And, you know, working with
- 12:41these standardization bodies and you know like what does a
- 12:44derivative look like and even if even if our data looks slightly
- 12:47different than our clients. At least we could standardize on
- 12:52how we talk about those those things.
- 12:56And that's what the legend platform, really excels at.
- 12:59So it's really a contribution to the broader data science data
- 13:04and analytics Community essentially.
- 13:06Ultimately absolutely absolutely.
- 13:08I think the platform is Is a general purpose platform for
- 13:12working with data and then you know we specialize on some
- 13:15standards and things like that working which is I think another
- 13:19contribution to the into the financial industry in general.
- 13:22Is just how do you get to that interop and standardization of
- 13:25at least data contracts or how we talk about different terms
- 13:28and relationships of different pieces of information.
- 13:32Please subscribe to our YouTube channel and hit the Subscribe
- 13:36button at the top of our website.
- 13:39We have Have a number of questions on Twitter so why
- 13:43don't we jump over to some of those?
- 13:46And we have an interesting question from our Salon con are
- 13:49salons, a regular listener and so thank you our salon for, for
- 13:53your, for listening, and for this great question.
- 13:56He says, given that data is such an important asset.
- 14:01How do you assign the financial value to that data?
- 14:05Who is it that? Decides that what data?
- 14:09As Financial value and how much it's worth and, and I'll just
- 14:14add to that. I'm assuming that, that helps
- 14:18that attachment of value is one of the things that guides your
- 14:22priorities in terms of where you focus your team focuses, what we
- 14:27do is we work with our business line, so I run our core data T.
- 14:31We work with our business lines about what their, what they need
- 14:36for their client to serve their clients better what they need
- 14:39for their business to scale what they need from us, to help them,
- 14:44sort of get an edge or innovate in the data realm.
- 14:48And so, a lot of it is about the value of the overall outcome
- 14:55versus just like ascribing value to some piece of data.
- 15:00And so, the way I, the way I think about from a core team,
- 15:05helping people is really what is the outcome?
- 15:09We're trying. To drive for our clients and for
- 15:12our businesses and for Innovation and really, look at
- 15:16that holistically. I think, again, I think it's a
- 15:19little bit of a, little bit of a misnomer, a mistake to sort of,
- 15:24try to say, okay, this dataset creates this much value, this
- 15:28creates that much value is like, what did we do for that business
- 15:31and for our clients and really we take those wins.
- 15:36As platform and data, winds collectively with, with those
- 15:40teams that makes perfect sense because the point is not some
- 15:46piece of data, the point or body of data.
- 15:49The point is, what are we doing with that data?
- 15:54And as you said, what are the resulting business outcomes and
- 15:59then you have a framework for for valuing, the the data
- 16:03because we're trying to get to To the outcome.
- 16:06Sorry, I didn't mean to answer for you.
- 16:07No, no, that's exactly right. That's it.
- 16:09That's a great. That's a great.
- 16:10Exactly. This, the right summary.
- 16:12And that's exactly how we do. Our oh, I like our return on
- 16:15investment. On the work we do, is we
- 16:18actually work hand in hand with those teens and say, okay, like
- 16:24it, we were we able to reduce Risk by X or were we able to
- 16:28help our clients do? Why better or faster or get them
- 16:33into better position for y or did we?
- 16:35Would enable the business to do some new thing that they were
- 16:39not able to do. And that's really like the ROI
- 16:43calc and at that level we have another question from Twitter.
- 16:48And another great question, this is from Natalie Bean who says,
- 16:54how do you balance data quality and data quantity?
- 16:58And we haven't even spoken about data quality yet, so this is a
- 17:02great question. When you think about data as a
- 17:05first-class, Class concept, when you think about data is the,
- 17:08when you do the same things, you do with your code, that you do
- 17:11with data, like, think about your data architecture up front,
- 17:14think about how you structure your data, how it relates to
- 17:17other pieces of data, right? And you do that, work up front,
- 17:21we have seen the yeah, The Upfront work takes a little bit
- 17:24longer but the, but the huge benefits of that become apparent
- 17:29as you're trying to scale these things.
- 17:31So we deal with it. We deal with the scale and the
- 17:35volume and the complexity, like any engineering organ.
- 17:40Is we build tools. We build platforms, and then we
- 17:43make sure that those scale to those problems.
- 17:46And right and so now when we do data Goldman Sachs it's like we
- 17:52do it on the platform and everyone knows like okay now the
- 17:56quants are going to get their data you know and days instead
- 17:59of months because we have set up the right platform constructs
- 18:03and Engineering constructs for that.
- 18:05So again it's not a perfect Silver Bullet, answer of like as
- 18:11things grow like, you know, there's some there's some
- 18:14equation but we have Seen that the investment in the platform
- 18:19and the tools and the workflows. That's that's the thing that
- 18:22helps us scale again. I find it really fascinating
- 18:26that the platform and as you said, the, the workflows play
- 18:30such a crucial role, but I suppose it's entirely logical
- 18:37that when you need that data when that data is so important
- 18:41and it needs to be right? And it needs to be consistently,
- 18:44write that Need that infrastructure and the
- 18:48automation to make it happen at that, at that level of quality.
- 18:52That's exactly right. This is my personal view coming
- 18:55from an engineering background. I solve pry eyes off, scale,
- 18:59problems of all your problems and complexity Problems by
- 19:02breaking it down into engineering steps and
- 19:04Engineering platforms. And so that's that's sort of
- 19:07been our ethos of this whole legend platform is really about
- 19:11that. How to attack that problem from
- 19:13an engineering lens. We have another Another
- 19:16excellent question from Twitter. This is from Lizabeth Shaw, who
- 19:21is now alluding to governance which you mentioned earlier, but
- 19:26we definitely should talk about that.
- 19:28And she says, how do you build in the mechanisms that support
- 19:32data policies? The point of the platform isn't
- 19:36that? It just magically makes these
- 19:38things go away. It's that it makes you think of
- 19:41them upfront, it makes you think of these Concepts.
- 19:45As you're doing your data design and your data work and so really
- 19:51when you when you're on the legend platform, thinking about
- 19:54your data, the first thing you think is what is your data
- 19:56contract, right? That's actually the first bit of
- 20:01the workflow is describe your data, describe how you want to
- 20:04publish it, describe how you want other people to see it or
- 20:07consume it how you want the track things like lineage.
- 20:14All of that is sort of built in baked in, as a first class
- 20:17concept. So again, I don't want to
- 20:19oversell the thing. You still have to think about
- 20:21it, you still have to do it, but the point is, bringing it up
- 20:24front, instead of hiding it as some secondary thing.
- 20:29And so, have to think about the entitlements, the security, the
- 20:32encryption, all of that. You think of, as sort of the
- 20:35first class assets, as you're just a first-class Concepts as
- 20:38you're designing, your data flow or your data workflow, or your
- 20:42data, Production or data consumption patterns, and all
- 20:47that sort of comes together in in the platform.
- 20:51It's fascinating again to me because it seems that your
- 20:56emphasis is definitely placed higher on that infrastructure
- 21:02and platforms. Then other Chief data officer,
- 21:06others that I've spoken with, but at the same time, you were
- 21:13Dealing with a level of data complexity combined with the
- 21:16speed and the and the financial consequences associated with it,
- 21:21that I think few other companies would have that set of
- 21:26combination of circumstances. Exactly like the quality the
- 21:30data has to be right. It has to be consistently,
- 21:33right, as you said, and the ramifications of that are pretty
- 21:39or pretty big. We have another question
- 21:42relating to to governance. And this is from again, from
- 21:46Arsenal and con. And he says, how do you decide
- 21:51what data is good to use? What is not?
- 21:55How do you address bias? He's in the data you're
- 21:58collecting and using. And this is interesting, he
- 22:03says, do the business lines, agree with your conclusions.
- 22:10So how do you also get everybody on the same page around this
- 22:12stuff? If we are not some like isolated
- 22:15team in the corner with like pointy hats, right?
- 22:19Like doing this in isolation. I think that's the first most
- 22:22important thing to get. Like, we are hands on keyboard
- 22:27together with the businesses, making sure that the data were
- 22:31using first is right, right? But the right for them and write
- 22:35for the use case, And really is solving their problem.
- 22:40So, the first point I would make is, you know, aligning ourselves
- 22:43with business outcomes and with the businesses is the first
- 22:47thing. So like not being in some
- 22:49isolated back room, like oh, we know best about everything,
- 22:53every piece of data. And again, the the financial
- 22:57domain is so complex that we would never even pretend to do
- 23:01that, right? You know, we have people who are
- 23:04highly skilled and finance, Financial areas.
- 23:07So that we could make those decisions with our our
- 23:11businesses, like, in sort of a joint venture fashion, but we
- 23:19would never say we're the experts in everything.
- 23:21So I think that's the first piece is like, you have to be
- 23:24connected to the business and the business outcomes.
- 23:26So that's that's the first piece then, then again the quality and
- 23:32the governance, right is then becomes an aligning incentives.
- 23:36Of a joint sort of venture, right?
- 23:38It's becomes there's a healthy tension, right?
- 23:40Of course we want to do things at scale and they want to sort
- 23:44of solve the problem immediately.
- 23:47But the point is again that, you know, bringing these two teams
- 23:50together really helps accelerate accelerate that and then get to
- 23:54the right answers and the right data.
- 23:56So this then leads to the question about the composition
- 24:01of your team. It's obvious that you have very
- 24:06three deep technology and data expertise, but as you just
- 24:12alluded in order to do your job. Again, I have to assume that you
- 24:18need equivalent Financial depth of expertise especially when you
- 24:24go into Concepts like derivatives, as you were, as you
- 24:27were describing earlier, we do have a team of stress or quants
- 24:32in the maybe more financial world, who Come from being on
- 24:38the desk. Understanding how data is used
- 24:42on the desks and have sort of a stem background, whether its
- 24:47technology math, physics, whatever who are equivalent,
- 24:51counterparties in Mighty counterparts, in my team that
- 24:54work with these groups and actually understand the
- 24:59financial domain, but cross over to the technical domain.
- 25:02So that's that's what we call our data design and curation
- 25:06team. I'm pretty pretty cool,
- 25:08high-powered, high-powered, crew, that that really is bound
- 25:12to those business. Was those folks which team are
- 25:16they part of? Are they part of your
- 25:18organization or part of the finance or trading organization?
- 25:23Where do they actually fit in our team?
- 25:26We have a subset of that team, right?
- 25:29But those, those sort of people are also all embedded in the
- 25:34business as well. So there's Bedded what we call
- 25:37embedded, death strats or embedded quants in various
- 25:41businesses. But then I have a sort of small
- 25:45selection of that team that works specifically on data with
- 25:51those other business teams, if that makes sense.
- 25:53What are the elements that comprise a successful data team
- 25:59such as you have? It really follows sort of the
- 26:02structure of the team like the platform team is high-powered
- 26:06soft. For engineers right there.
- 26:08They come from a software background.
- 26:10Their goal is to make the software bulletproof and build
- 26:13the right workflows for data engineers.
- 26:16And the second bucket is basically data, Engineers,
- 26:19people who are content experts but can use the platform
- 26:23configure. The platform build data
- 26:26pipelines, build curation pipelines.
- 26:30Build data models that then get shared out then the third
- 26:36bucket. Is really this sort of hybrid
- 26:40hybrid. Again, I'll use the word strap
- 26:41because that's what we use internally.
- 26:43This hybrid team, that's really straddles sort of deep finance
- 26:48and deep Tech together. And then the fourth is sort of
- 26:53our governance, you know, framework team.
- 26:56They're the people who sort of set the policies set the
- 27:00framework on and and and work with the divisions to sort of,
- 27:05make sure there. They're working in the bounds of
- 27:08our framework. One of the topics that we have,
- 27:11not really touched upon is the notion of data interoperability
- 27:18and I know that's important to you.
- 27:20So can you tell us about that? We touched a little bit of pain.
- 27:24It, that's sort of why a big big reason we open sourced.
- 27:28Our Legend platform was exactly for that was because we felt
- 27:33that like if we could bring some Platform standardization in the
- 27:38industry for us or clients are counterparties that then at
- 27:44least the discussions about how data works or how it should be
- 27:48connected or how it should be described in the industry that
- 27:52we could help push that forward and we can help push that
- 27:55forward with standards with other bodies, but they could all
- 28:00do that work in one sort of saying way in in our platform.
- 28:04Well, not it's not even our For me, in the open source Community
- 28:08platform and and that will benefit everybody.
- 28:13So it's a bit, it's absolutely a big, a big, big play, I'll
- 28:18mention, you know, one project we're doing in the in US
- 28:22Community which is the open source nonprofit that we gave
- 28:27gave Legend to. We brought is though, which is
- 28:30this the derivative standards d'oeuvres body?
- 28:34They have built a The model called CDM, common domain model
- 28:38for derivatives and now that sort of that data model is now
- 28:44available in legend for people to collaborate on in a
- 28:47collaborative environment out in the wild it's has nothing to do
- 28:50with Goldman Sachs they're just using the platform.
- 28:52So these are like pretty cool things that you know we're
- 28:57trying to do to push the interop and and standardization out how
- 29:01has the uptake bin in the broader community?
- 29:06Of of this platform and data. Interoperability the cool part
- 29:11is that our code is out there we have clients, actually deploying
- 29:15it using it. We have other counterparties
- 29:17also looking at it, doing pocs, in their own environment, but
- 29:21the flip side of that is, you know, building an open source
- 29:24Community is very difficult. I mean, we vastly underestimated
- 29:30how much time and effort and energy goes into, you know,
- 29:35making Making sure people know how to use the thing, the
- 29:38documentation. Do they understand the value?
- 29:40Can they contribute code back? Like, all of these things, we
- 29:45knew it would take time, but I think we sort of grossly
- 29:49underestimated how much time to build sort of that, at that big
- 29:53Community around these. These projects, what are the
- 29:57benefits that you have presented to the community as to why they
- 30:01should engage in this, it has helped us Date of better at
- 30:07Goldman, is helped us organize our data.
- 30:10It has helped break down silos. It has helped the data quality
- 30:13massively. It has helped data governance
- 30:15aspects, massively internally. And so we point to sort of those
- 30:20success stories about how we do it internally that we could also
- 30:24help clients do the same thing. And you know if there's appetite
- 30:30of course that we could also help the broader industry break
- 30:33down those silos as well. We have Another couple of
- 30:36questions that are coming in from Twitter.
- 30:39Do you think of your data or your platform as a product that
- 30:47you can sell, given that you've built all of this
- 30:50infrastructure? I'll say that's not the
- 30:52intention of doing it. I think, though, again, we have
- 30:57seen that as clients use it as other Banks, use it, that people
- 31:01sort of come back to us and ask, okay, well, what could we What
- 31:06could the support model B, or, hey, can you host this as a SAS
- 31:10for us? So that we don't have to deal
- 31:12with setting up the infrastructure, running the
- 31:14infrastructure or things like that.
- 31:15So I think we're still again, just not to oversell anything.
- 31:19I think we're still in a very crawl stages of even thinking
- 31:23about that community and how we could help do better and better
- 31:26but Our intention isn't like Like make money off of this
- 31:32thing. It's really about the
- 31:34standardization parts now if people see a large adoption and
- 31:38see a value, I think it could be an exciting potential
- 31:42opportunity. We have another question from
- 31:46Twitter. Can you give any examples where
- 31:50using the platform has resulted in innovative ways of for users
- 31:57to use and or combined? the data we sort of took again,
- 32:02a data-driven approach to helping our firm and our Risk
- 32:08Managers and our sales. People in our Traders,
- 32:11understand that risk from a data-driven perspective, right?
- 32:16We have combined all that data and information into our
- 32:21platform and have basically given I hate to use this word
- 32:25but democratized the access to that our team has linked a bunch
- 32:30of that data together done. Some interesting analysis but
- 32:34more importantly, we have In that those based sets of data,
- 32:38those relationships, the relationships between those data
- 32:40that may not be obvious. To our users hands.
- 32:44That now, they have the access to that information and now they
- 32:49could come up with creative things or risk, mitigating
- 32:52things to on top of that data. So I think that's been, you
- 32:57know, maybe one that resonates pretty recently here, it's kind
- 33:02of the reverse of Shadow it that cios didn't like years ago.
- 33:09Some going back, maybe, five years, you're Ali, taking the
- 33:13data and I assume the tools and placing it in the hands of
- 33:18users, so that they can be creative.
- 33:21They can innovate with with that data.
- 33:24Exactly. But in a fully governed quality
- 33:28way, right, where that shadow, it problem isn't really a
- 33:33problem. We have put all the right guard
- 33:35rails and all the right governance around that.
- 33:38So now people can actually innovate in a safe space.
- 33:43In a safe way but it's not just lock to our team.
- 33:47Can you give us some insight into what kind of governance or
- 33:52how do you balance the need for security for privacy against
- 33:58making that data, available accessible and easy to use for
- 34:02folks before we do anything with data.
- 34:05Right? We have first of all very strict
- 34:08policies and Frameworks about, you know, understanding who's
- 34:11allowed to see what. Data.
- 34:14And even if you're allowed to see it, should you see it?
- 34:17You know, we call it need to know, right?
- 34:18Like just because you're allowed to see it.
- 34:20Maybe you don't actually need to know that client information.
- 34:23So first of all, as an overarching theme, I'd say like
- 34:28one of the one of the most critical pieces is is you know
- 34:32writing those policies down and actually then sort of enforcing
- 34:37them in the platform. But before we do anything like
- 34:40that is a clear thing and then and we have very clear rules
- 34:44About who can get access to what?
- 34:46How, and then again, would be a little broken record encode,
- 34:51those rules in the platform and make sure that that is like a
- 34:55Slayer. Like if we get that wrong,
- 34:57nothing matters, right? So the first thing is to make
- 35:01sure that that Baseline Works before before giving access to
- 35:07him. So in other words, as long as
- 35:10the foundation of governance, A risk control, compliance with
- 35:17regulations, whatever is necessary.
- 35:19As long as those elements are in place, then you're able to share
- 35:26the data. And let people, I was going to
- 35:30say have free range, obviously that's not the case, but have
- 35:33enough rain, where they can use it.
- 35:36Use that data creatively in the service of whatever their
- 35:38business goals happen to be much better said, then that then
- 35:42white man eyes that it but Exactly, exactly.
- 35:45The right mental model. What advice do you have for
- 35:51folks who are building a data strategy, based on what you've
- 35:58learned at and done at Goldman Sachs?
- 36:01Make sure you attach yourself to business outcomes.
- 36:06The things that people care about at your company, right?
- 36:09That I think, As a technologist. You know sometimes as
- 36:14technologists in general maybe I'll make a broad generalization
- 36:17sometimes. You know, we care a lot more
- 36:19about the tech than the outcome but specifically for data
- 36:25strategy, I think it's even more important to over-index on the
- 36:28outcomes because data sort of becomes this nebulous thing
- 36:33where now you know like well what does it mean to have a data
- 36:35strategy? What is that like what is data
- 36:37even mean? Why do I need that?
- 36:39Like an abstract? I get why I want information,
- 36:43but like what are you talking about?
- 36:45So the first thing, I the first, first thing I talked about
- 36:48always is attached yourself, the business outcomes and show how
- 36:53the data and the data strategy actually makes those cheaper
- 36:58faster, better easier makes money for your clients.
- 37:02Helps reduce risk, help. Save money, do so.
- 37:06That's that's the first sort of Baseline advice.
- 37:11I give to everybody. And then then I think, you know,
- 37:17but then below that then become sort of okay set out.
- 37:22What is the platform strategy going to be?
- 37:25How are you going to actually make the engineering work?
- 37:28Think about the business workflows and the Pro workflows.
- 37:35And then, you know, we talked a little bit about the org
- 37:37structure and the framework. So I think those are also key
- 37:40pieces, right? Make sure that, you know, if
- 37:44you're going to do it in an engineering strategy, you have a
- 37:46strong platform team, if you're going to, you know, have a
- 37:51Content T. Make sure they actually
- 37:52understand the domain that you're working with.
- 37:59And so bringing those like or pieces together.
- 38:04But, but the number one key is like, make sure you're driving
- 38:09outcomes. What's the relationship between
- 38:11your data strategy, and Cloud, where does cloud fit into all of
- 38:15this? I can't believe we haven't
- 38:16spoken about that. I think of the cloud again, as a
- 38:19tool in the toolbox, right? It's not a means to an end it's
- 38:25in my Arsenal how I want to build my platform.
- 38:28Yeah. I want infinite scalability, I
- 38:31wanted it, I want, you know, other people have done, the hard
- 38:35work on the infrastructure front, I want great databases
- 38:39that people have built to not rebuild that on my own.
- 38:42And so to me, the cloud is a great enabler, it's a great
- 38:46tool, my toolbox, it's a great way to get the scale.
- 38:51So it's definitely a big part of the overall data strategy.
- 38:56But again, it just to be clear. I just make sure like, It's not
- 39:01a thing for the sake of a thing, it's like, okay.
- 39:05These are great components, great pieces of infrastructure
- 39:08that other people that I can now stand on the shoulder of giants
- 39:12execute on. And that's the way I think about
- 39:15it. It was super important part of
- 39:16the strategy just to be clear. I can see that you are you're
- 39:21very purpose-driven, you are always coming back to the
- 39:25reference point. Why are we doing it?
- 39:26What are we doing at? What are we getting out of it
- 39:28from them from the business? Outcome standpoint.
- 39:30That's Very, very, very clear. And finally, one last question,
- 39:36what is the relationship between data and building business
- 39:40models economic models Financial models?
- 39:44How does how to those pieces fit together in technology?
- 39:47It comes down to algorithms and data and so like those are the
- 39:51two big inputs, right? Like, you need the data to be
- 39:54great. You need it to be clean and
- 39:56needed to be organized. You need to make sure that like
- 39:59all those pieces are set, it's easy.
- 40:01Accessible. It's findable to govern and then
- 40:05that becomes a major input into the Algos, whether you're doing
- 40:10forecasting or algorithmic trading or helping your client
- 40:14with something that those pieces just have to like work together
- 40:19as a as a team to get the job done.
- 40:24Again, I'm putting words in your mouth, but this is just
- 40:26summarizing what you've just been talking about is the
- 40:29linkage to the out. Outcome and being clear about,
- 40:32what's the data going in? What's the expected result at
- 40:34the other and and ensuring that the to match up?
- 40:38Yep, yep, exactly. And with that, unfortunately,
- 40:42we're out of time Nema. I just want to say a huge.
- 40:45Thank you for spending time with us.
- 40:47I really, really appreciate it. Thank you, thanks for the great
- 40:50questions. Thanks audience, amazing
- 40:53questions. I really had a really fun time
- 40:55talking to talking to you Michael.
- 40:57So thanks for having me, everybody.
- 41:00Thank you. For watching, I just want to say
- 41:03a huge thank you to Nema Rafael. He is the head of global data
- 41:09engineering and the chief data officer of Goldman Sachs.
- 41:15Now before you go, please subscribe to our YouTube channel
- 41:20and hit the Subscribe button at the top of our website.
- 41:24Actually, you know, the Subscribe button has moved to
- 41:26the bottom of our website, so hit the Subscribe button at the
- 41:29bottom of our website. So We can send you our
- 41:31newsletter and keep you up-to-date on our upcoming live
- 41:34shows. Thanks so much, everybody.
- 41:36Hope you have a great day and we'll see you next time.