1 00:00:00,000 --> 00:00:04,200 Today we are discussing data at Goldman Sachs. 2 00:00:04,600 --> 00:00:08,900 We're speaking with Nemo, Raphael the firm's head of data 3 00:00:08,900 --> 00:00:11,700 engineering and their Chief data officer. 4 00:00:11,700 --> 00:00:15,300 I've been at Bowman about 20 years, always as a software 5 00:00:15,300 --> 00:00:20,400 engineer data, engineer rat, or want as we call volunteer 6 00:00:20,400 --> 00:00:23,200 stress. And now, as you mentioned, I run 7 00:00:23,200 --> 00:00:25,900 our data team or data, engineering team and chief data 8 00:00:25,900 --> 00:00:27,800 officer. For Goldman data engineering, 9 00:00:27,800 --> 00:00:32,200 Goldman is probably a little little bit different than other 10 00:00:32,200 --> 00:00:36,400 places. I, I, sort of organize my brain 11 00:00:36,400 --> 00:00:39,200 and my aim and sort of three buckets. 12 00:00:39,200 --> 00:00:42,500 We have our platform team, which hopefully, I'll have a little 13 00:00:42,500 --> 00:00:46,300 more time to talk about and our curation, our content and 14 00:00:46,300 --> 00:00:51,000 curation team. And also as a chief data 15 00:00:51,000 --> 00:00:54,200 officer, with my chief data officer hat on sort of our 16 00:00:54,500 --> 00:00:57,800 governance and quality teams. So that's that's sort of the 17 00:00:57,800 --> 00:01:03,400 background of the work we do. Data is so obviously important 18 00:01:03,400 --> 00:01:08,100 to financial services but it would be really interesting to 19 00:01:08,100 --> 00:01:14,500 hear your perspective on the role of data how data fits into 20 00:01:14,500 --> 00:01:18,100 the world of Goldman Sachs information is the lifeblood of 21 00:01:18,100 --> 00:01:20,700 financial services and people say well what is that? 22 00:01:20,900 --> 00:01:24,700 And and I think you know, in financial services at Goldman, 23 00:01:24,700 --> 00:01:30,400 of course you know information is is actually our currency So, 24 00:01:30,400 --> 00:01:34,900 a lot of a lot of decision making a lot of helping our 25 00:01:34,900 --> 00:01:40,600 clients, a lot of innovation is all based on what information we 26 00:01:40,600 --> 00:01:44,600 see out in the world and what information we see internally. 27 00:01:46,400 --> 00:01:49,700 And and so organizing that making sure it's readily 28 00:01:49,700 --> 00:01:54,400 accessible, making sure people could get to that data quickly 29 00:01:55,400 --> 00:01:59,500 to, to do to do their work or to help their clients. 30 00:02:00,200 --> 00:02:05,100 Starts becoming sort of like a core, a core component of doing 31 00:02:05,100 --> 00:02:08,500 business, that Goldman Sachs, you mentioned, the various 32 00:02:08,500 --> 00:02:11,600 aspects of your role, how you break things up. 33 00:02:11,600 --> 00:02:13,800 Do you want to dive into that? And tell us a little bit more 34 00:02:13,800 --> 00:02:16,100 about that. The first thing I think about 35 00:02:16,100 --> 00:02:20,100 data is that it has to be a first class asset. 36 00:02:20,500 --> 00:02:23,100 Right at Goldman. People have to believe that it 37 00:02:23,100 --> 00:02:26,300 could help solve problems. Help, our clients help us 38 00:02:26,300 --> 00:02:31,800 innovate and financial services. So, so the first thing we sort 39 00:02:31,800 --> 00:02:35,700 of did is okay, how do we want to work with data Goldman and 40 00:02:35,700 --> 00:02:39,600 again, as an engineer, that always push me to sort of think 41 00:02:39,600 --> 00:02:44,200 about data as as an engineer would think about their code. 42 00:02:44,200 --> 00:02:47,000 So I the analogy, I always say, or maybe it's not even analogy, 43 00:02:47,000 --> 00:02:51,400 but the way I always talk about data is like, you got to think 44 00:02:51,400 --> 00:02:53,800 about data, just like you were thinking about your code and 45 00:02:53,800 --> 00:02:57,500 that's sort of has pushed us to build a platform team. 46 00:02:58,200 --> 00:03:02,000 That really thinks about that. Things about data engineering as 47 00:03:02,000 --> 00:03:05,200 an engineering function. He thinks about the workflows of 48 00:03:05,200 --> 00:03:10,200 data Engineers, how we could help get, leverage and scale to 49 00:03:10,200 --> 00:03:13,000 our businesses so that they could start making decisions 50 00:03:13,000 --> 00:03:17,200 better faster, cheaper easier. And so the platform team and my 51 00:03:17,200 --> 00:03:20,900 team is really laser focused on that, right? 52 00:03:21,000 --> 00:03:24,200 How do you make data engineering workflows just seamless? 53 00:03:24,400 --> 00:03:29,600 Beautiful, and all the things that, you know, Developers are 54 00:03:29,600 --> 00:03:33,800 used to right? I ve s code completion. 55 00:03:35,300 --> 00:03:39,000 Making sure that we have like reproducibility of our data. 56 00:03:39,000 --> 00:03:43,500 Like we have code think that so all of that sort of becomes 57 00:03:43,500 --> 00:03:48,500 their first class like role is that won't is to only think 58 00:03:48,500 --> 00:03:53,400 about how do we make data awesome working with data also 59 00:03:54,500 --> 00:03:58,800 and then we have a Content team. So again like I said, A lot of a 60 00:03:58,800 --> 00:04:03,100 lot of our workflows at Goldman run on data. 61 00:04:03,100 --> 00:04:07,800 So you can imagine, you know, part of our team is a real-time 62 00:04:07,800 --> 00:04:09,900 Market data team. So all the all the data 63 00:04:09,900 --> 00:04:12,700 streaming from exchanges or different, venues things like 64 00:04:12,700 --> 00:04:15,600 that in real time, super low latency. 65 00:04:15,600 --> 00:04:19,899 You know, we're talking about like millisecond latency here 66 00:04:19,899 --> 00:04:23,600 right to get to Trader screens or to get into the Algos. 67 00:04:23,800 --> 00:04:26,800 And so there's a, there's a whole team of mine which is in 68 00:04:26,800 --> 00:04:30,400 the curation business. The content business, whether 69 00:04:30,400 --> 00:04:32,800 it's real-time Market data or things. 70 00:04:32,800 --> 00:04:36,800 Like our reference data that we use a, you know, like who our 71 00:04:36,800 --> 00:04:39,400 clients or what products, can we trade things like that? 72 00:04:39,400 --> 00:04:44,700 And so that team is really, I'd say, sort of a data engineering 73 00:04:44,700 --> 00:04:47,800 bent team that uses our platforms to sort of do their 74 00:04:47,800 --> 00:04:52,400 work, make sure the data is amazing, quality wise accessible 75 00:04:54,200 --> 00:04:58,000 and ready to go. And then, the third bit is Is 76 00:04:58,008 --> 00:05:02,700 really about the framework that we do data governance and data 77 00:05:02,700 --> 00:05:08,400 quality around like so a little bit around policy and Frameworks 78 00:05:08,400 --> 00:05:11,700 but then how do we push our platform team to then build in 79 00:05:11,700 --> 00:05:15,900 those mechanisms or controls on the platform. 80 00:05:15,900 --> 00:05:18,500 So when you do your work at Goldman, you sort of get those 81 00:05:18,500 --> 00:05:20,600 for free, I say or complimentary. 82 00:05:20,600 --> 00:05:24,800 So that's sort of the, a little more about the three buckets, we 83 00:05:24,800 --> 00:05:27,900 talked about here. What I find particularly 84 00:05:28,000 --> 00:05:32,700 Interesting is unlike most businesses which are of course 85 00:05:32,700 --> 00:05:39,000 have their their business data. You're dealing with huge volumes 86 00:05:39,000 --> 00:05:45,800 of real-time information of real-time data at which there is 87 00:05:45,800 --> 00:05:50,700 enormous amounts of money and risk at stake. 88 00:05:51,000 --> 00:05:56,700 I have to imagine that that places and additional kind of 89 00:05:56,700 --> 00:06:00,100 intense bird. Ian on you, as you're thinking 90 00:06:00,100 --> 00:06:04,100 about this whole technology landscape, I would sort of flip 91 00:06:04,100 --> 00:06:06,600 it in a positive. I don't think it's actually, we 92 00:06:06,600 --> 00:06:09,000 don't think of it as a burden. We think of it as a real 93 00:06:09,000 --> 00:06:12,500 opportunity, right? That I think, is the first sort 94 00:06:12,500 --> 00:06:14,600 of mindset shift. You have to think about when 95 00:06:14,600 --> 00:06:16,400 you're in these sort of high-stakes games. 96 00:06:16,400 --> 00:06:17,700 Yeah, of course, it's high stakes. 97 00:06:17,700 --> 00:06:21,000 Of course, things could go wrong, but really, it's more 98 00:06:21,000 --> 00:06:24,100 about the opportunities to make sure that again, we're helping 99 00:06:24,100 --> 00:06:27,300 our clients, do the right thing, we're helping the economy, we're 100 00:06:27,300 --> 00:06:29,700 hoping the Financial system do the right thing. 101 00:06:29,700 --> 00:06:34,200 So we sort of take pride in that instead of a sort of Burden 102 00:06:34,200 --> 00:06:37,000 thing. And the other one thing I'd say 103 00:06:37,000 --> 00:06:43,200 is, I actually like to say you said, huge amounts of data, I 104 00:06:43,200 --> 00:06:46,200 actually like to say Goldman Sachs were in the sort of medium 105 00:06:46,200 --> 00:06:50,500 sized data but very complex data World. 106 00:06:50,500 --> 00:06:54,000 Our data super complex, the types of data are complex, the 107 00:06:54,000 --> 00:06:57,700 speed is complex, there's everything from super real. 108 00:06:57,900 --> 00:07:01,300 Time low latency all the way to, you know, end of day batch 109 00:07:01,300 --> 00:07:05,100 processing the relationships between our data as complex, 110 00:07:06,300 --> 00:07:10,100 right. The product catalog is complex, 111 00:07:10,100 --> 00:07:13,400 right? It's not just use of whatever 112 00:07:14,000 --> 00:07:17,700 its actual financial instruments that people have have made up. 113 00:07:18,500 --> 00:07:22,300 So the complexity is really the really interesting and cool 114 00:07:22,300 --> 00:07:29,100 challenge here versus I'd say the volume what makes The data 115 00:07:29,100 --> 00:07:31,200 complex, or just describing a little bit. 116 00:07:31,200 --> 00:07:34,500 Can you drill into that? The complexity again comes from 117 00:07:34,500 --> 00:07:37,700 various angles, like you mentioned the complexity comes 118 00:07:37,700 --> 00:07:41,700 from the, from the speed is one angle one dimension, right? 119 00:07:41,700 --> 00:07:43,600 But also the relationships of the data. 120 00:07:43,600 --> 00:07:49,500 So you could imagine, you know, there's data about the stocks 121 00:07:49,500 --> 00:07:54,000 and bonds that peep that our clients want to trade for one 122 00:07:54,000 --> 00:07:57,600 example, but then there's a whole infrastructure. 123 00:07:57,900 --> 00:08:03,100 Around those like I don't know how many people here are 124 00:08:03,400 --> 00:08:05,500 familiar with financial services, but there's 125 00:08:05,500 --> 00:08:07,600 derivatives on those products, right? 126 00:08:07,600 --> 00:08:09,700 So now you're not just talking about stocks and bonds, but 127 00:08:09,700 --> 00:08:12,300 you're talking about really complex. 128 00:08:12,800 --> 00:08:19,900 Now, algorithms plus complex data elements about now that 129 00:08:19,900 --> 00:08:24,200 have a relationship that have been now made up in the world on 130 00:08:24,200 --> 00:08:27,500 those stocks and bonds so layering of this data and the 131 00:08:27,900 --> 00:08:30,100 Alexa tea. Now that people have come up 132 00:08:30,100 --> 00:08:33,200 with such creative ideas to help our clients and things like that 133 00:08:34,200 --> 00:08:38,100 starts pushing the boundaries of like how this data is related 134 00:08:38,100 --> 00:08:43,100 and interconnected and curated. Hopefully, that gives a little 135 00:08:43,100 --> 00:08:48,100 bit of a sense. You are managing these very 136 00:08:48,100 --> 00:08:55,100 large volumes of real-time data. Do you want to tell us at all 137 00:08:55,100 --> 00:08:59,000 about your, your technology infrastructure? 138 00:08:59,000 --> 00:09:02,800 It's not a question. I usually ask, but it seems that 139 00:09:02,800 --> 00:09:08,400 in this case given the the equities, the derivatives, 140 00:09:08,400 --> 00:09:10,900 everything else that you had just mentioned it. 141 00:09:11,100 --> 00:09:13,800 Seems the technology infrastructure has to play a 142 00:09:13,800 --> 00:09:16,000 very important role in the structure. 143 00:09:16,000 --> 00:09:20,200 Absolutely plays a huge role. Here, we have an incredible 144 00:09:20,200 --> 00:09:24,100 infrastructure team and you could imagine again that this 145 00:09:24,100 --> 00:09:27,600 goes all the way down to sort of the hardware Hardware layers, 146 00:09:27,600 --> 00:09:34,800 like the networking stack, the computers you use the network 147 00:09:34,800 --> 00:09:40,300 cards in the computers use. All this stuff sort of actually 148 00:09:40,300 --> 00:09:44,800 matters at At this sort of latency in scale, but I got it. 149 00:09:45,000 --> 00:09:47,600 I don't want a super over index on the Real Time stuff. 150 00:09:47,600 --> 00:09:51,800 The Real Time stuff is seen as a very important, very critical 151 00:09:51,800 --> 00:09:57,000 and important part of of of our of our world, for sure. 152 00:09:57,800 --> 00:10:00,700 And the team does an amazing job all the way from the hardware 153 00:10:00,700 --> 00:10:05,700 layer up. But we also have, you know, data 154 00:10:05,700 --> 00:10:09,200 challenges just after, you know, as the real-time data. 155 00:10:11,000 --> 00:10:14,100 As in, you know, when you do a trade right, that store also 156 00:10:14,100 --> 00:10:18,100 starts being complex and so we have built sort of a data 157 00:10:18,100 --> 00:10:23,400 platform here, we call Legend and we've actually open source 158 00:10:23,400 --> 00:10:25,000 that recently in the last couple of years. 159 00:10:25,000 --> 00:10:27,900 We've built this Tech stack over the last 10 years. 160 00:10:27,900 --> 00:10:31,700 Internally, two years about two years ago, maybe three. 161 00:10:31,700 --> 00:10:35,400 Now we open sourced it it's fully on GitHub, you can check 162 00:10:35,400 --> 00:10:38,500 it out. We gave the code to a non-profit 163 00:10:38,500 --> 00:10:43,500 and open source nonprofit called Fitness, Around GitHub happy to 164 00:10:43,500 --> 00:10:45,600 talk more about about that as well. 165 00:10:46,000 --> 00:10:49,900 Why did you open source at first glance? 166 00:10:50,200 --> 00:10:53,100 It seems like an odd decision to me. 167 00:10:53,100 --> 00:10:57,500 Because one would think that a firm like Goldman Sachs would 168 00:10:57,500 --> 00:11:00,800 want to keep all of that infrastructure to yourself. 169 00:11:01,200 --> 00:11:03,900 There's a couple of reasons. One is, as we were building 170 00:11:03,900 --> 00:11:08,200 this, we talked to a lot of clients who are having exactly 171 00:11:08,200 --> 00:11:12,800 the same data challenges. We Were and, you know, we're 172 00:11:12,800 --> 00:11:17,500 very obviously, a very client-centric firm. 173 00:11:17,500 --> 00:11:20,800 So we thought look, this has helped us so much internally. 174 00:11:20,800 --> 00:11:24,100 This platform has helped us so much internally, we'd love to 175 00:11:24,100 --> 00:11:27,100 give it a chance. Help our clients to help push 176 00:11:27,100 --> 00:11:31,200 the industry forward as well. And and so you also have to 177 00:11:31,200 --> 00:11:34,400 separate a little bit about the platform itself versus the 178 00:11:34,400 --> 00:11:37,500 content that we curate and work in the platform. 179 00:11:37,500 --> 00:11:42,700 So, we haven't yet. Or maybe never will open source 180 00:11:42,700 --> 00:11:47,300 the actual content, but we could talk about that as well. 181 00:11:47,900 --> 00:11:50,400 But the platform and the work we've done to sort of 182 00:11:50,400 --> 00:11:53,100 standardize, how we work with data. 183 00:11:53,100 --> 00:11:58,500 We thought that was so powerful. That that actually giving it out 184 00:11:58,500 --> 00:12:00,900 to the community and building a community around, that could 185 00:12:00,900 --> 00:12:05,200 actually help the whole industry and our clients and and 186 00:12:05,200 --> 00:12:09,800 ourselves you are wanting to share best practices. 187 00:12:10,000 --> 00:12:14,100 Is approaches to working with data things like that with the 188 00:12:14,100 --> 00:12:17,400 community and then of course, everybody's going to have their 189 00:12:17,400 --> 00:12:20,500 own data content within within those constructs. 190 00:12:20,900 --> 00:12:23,300 That's exactly right. The interesting part is even the 191 00:12:23,500 --> 00:12:26,100 so even if we don't give out the content, right? 192 00:12:26,900 --> 00:12:31,800 We are working with industry standard bodies, to even 193 00:12:31,800 --> 00:12:35,100 describe the data and structure the data, and the linkages to 194 00:12:35,100 --> 00:12:38,000 the data. So, it's a really cool and 195 00:12:38,000 --> 00:12:41,600 Powerful technique. And, you know, working with 196 00:12:41,600 --> 00:12:44,400 these standardization bodies and you know like what does a 197 00:12:44,408 --> 00:12:47,900 derivative look like and even if even if our data looks slightly 198 00:12:47,900 --> 00:12:52,700 different than our clients. At least we could standardize on 199 00:12:52,700 --> 00:12:56,200 how we talk about those those things. 200 00:12:56,200 --> 00:12:59,300 And that's what the legend platform, really excels at. 201 00:12:59,700 --> 00:13:04,100 So it's really a contribution to the broader data science data 202 00:13:04,100 --> 00:13:06,300 and analytics Community essentially. 203 00:13:06,300 --> 00:13:08,300 Ultimately absolutely absolutely. 204 00:13:08,300 --> 00:13:12,300 I think the platform is Is a general purpose platform for 205 00:13:12,300 --> 00:13:15,700 working with data and then you know we specialize on some 206 00:13:15,700 --> 00:13:19,100 standards and things like that working which is I think another 207 00:13:19,100 --> 00:13:22,000 contribution to the into the financial industry in general. 208 00:13:22,000 --> 00:13:24,900 Is just how do you get to that interop and standardization of 209 00:13:25,600 --> 00:13:28,700 at least data contracts or how we talk about different terms 210 00:13:28,700 --> 00:13:32,000 and relationships of different pieces of information. 211 00:13:32,400 --> 00:13:36,700 Please subscribe to our YouTube channel and hit the Subscribe 212 00:13:36,700 --> 00:13:39,300 button at the top of our website. 213 00:13:39,300 --> 00:13:43,300 We have Have a number of questions on Twitter so why 214 00:13:43,300 --> 00:13:45,700 don't we jump over to some of those? 215 00:13:46,300 --> 00:13:49,900 And we have an interesting question from our Salon con are 216 00:13:49,900 --> 00:13:53,400 salons, a regular listener and so thank you our salon for, for 217 00:13:53,400 --> 00:13:55,600 your, for listening, and for this great question. 218 00:13:56,100 --> 00:14:00,800 He says, given that data is such an important asset. 219 00:14:01,000 --> 00:14:05,600 How do you assign the financial value to that data? 220 00:14:05,600 --> 00:14:09,700 Who is it that? Decides that what data? 221 00:14:09,900 --> 00:14:14,500 As Financial value and how much it's worth and, and I'll just 222 00:14:14,500 --> 00:14:17,900 add to that. I'm assuming that, that helps 223 00:14:18,200 --> 00:14:22,900 that attachment of value is one of the things that guides your 224 00:14:22,900 --> 00:14:27,300 priorities in terms of where you focus your team focuses, what we 225 00:14:27,300 --> 00:14:31,200 do is we work with our business line, so I run our core data T. 226 00:14:31,200 --> 00:14:35,600 We work with our business lines about what their, what they need 227 00:14:36,000 --> 00:14:39,500 for their client to serve their clients better what they need 228 00:14:39,500 --> 00:14:44,100 for their business to scale what they need from us, to help them, 229 00:14:44,100 --> 00:14:48,200 sort of get an edge or innovate in the data realm. 230 00:14:48,800 --> 00:14:54,700 And so, a lot of it is about the value of the overall outcome 231 00:14:55,100 --> 00:15:00,600 versus just like ascribing value to some piece of data. 232 00:15:00,700 --> 00:15:05,600 And so, the way I, the way I think about from a core team, 233 00:15:05,600 --> 00:15:09,400 helping people is really what is the outcome? 234 00:15:09,400 --> 00:15:12,600 We're trying. To drive for our clients and for 235 00:15:12,700 --> 00:15:16,700 our businesses and for Innovation and really, look at 236 00:15:16,700 --> 00:15:19,100 that holistically. I think, again, I think it's a 237 00:15:19,108 --> 00:15:24,900 little bit of a, little bit of a misnomer, a mistake to sort of, 238 00:15:24,900 --> 00:15:28,500 try to say, okay, this dataset creates this much value, this 239 00:15:28,500 --> 00:15:31,900 creates that much value is like, what did we do for that business 240 00:15:31,900 --> 00:15:35,200 and for our clients and really we take those wins. 241 00:15:36,000 --> 00:15:40,600 As platform and data, winds collectively with, with those 242 00:15:40,600 --> 00:15:46,400 teams that makes perfect sense because the point is not some 243 00:15:46,400 --> 00:15:49,500 piece of data, the point or body of data. 244 00:15:49,500 --> 00:15:53,800 The point is, what are we doing with that data? 245 00:15:54,200 --> 00:15:59,300 And as you said, what are the resulting business outcomes and 246 00:15:59,300 --> 00:16:03,600 then you have a framework for for valuing, the the data 247 00:16:03,900 --> 00:16:06,300 because we're trying to get to To the outcome. 248 00:16:06,400 --> 00:16:07,800 Sorry, I didn't mean to answer for you. 249 00:16:07,900 --> 00:16:09,400 No, no, that's exactly right. That's it. 250 00:16:09,400 --> 00:16:10,700 That's a great. That's a great. 251 00:16:10,700 --> 00:16:12,600 Exactly. This, the right summary. 252 00:16:12,600 --> 00:16:15,900 And that's exactly how we do. Our oh, I like our return on 253 00:16:15,908 --> 00:16:18,400 investment. On the work we do, is we 254 00:16:18,400 --> 00:16:23,900 actually work hand in hand with those teens and say, okay, like 255 00:16:24,200 --> 00:16:28,500 it, we were we able to reduce Risk by X or were we able to 256 00:16:28,500 --> 00:16:33,000 help our clients do? Why better or faster or get them 257 00:16:33,000 --> 00:16:35,400 into better position for y or did we? 258 00:16:35,800 --> 00:16:39,400 Would enable the business to do some new thing that they were 259 00:16:39,400 --> 00:16:43,100 not able to do. And that's really like the ROI 260 00:16:43,100 --> 00:16:48,700 calc and at that level we have another question from Twitter. 261 00:16:48,700 --> 00:16:53,900 And another great question, this is from Natalie Bean who says, 262 00:16:54,000 --> 00:16:58,900 how do you balance data quality and data quantity? 263 00:16:58,900 --> 00:17:02,000 And we haven't even spoken about data quality yet, so this is a 264 00:17:02,000 --> 00:17:05,000 great question. When you think about data as a 265 00:17:05,000 --> 00:17:07,800 first-class, Class concept, when you think about data is the, 266 00:17:08,400 --> 00:17:11,099 when you do the same things, you do with your code, that you do 267 00:17:11,099 --> 00:17:14,599 with data, like, think about your data architecture up front, 268 00:17:14,700 --> 00:17:17,900 think about how you structure your data, how it relates to 269 00:17:17,900 --> 00:17:21,400 other pieces of data, right? And you do that, work up front, 270 00:17:21,400 --> 00:17:24,599 we have seen the yeah, The Upfront work takes a little bit 271 00:17:24,599 --> 00:17:29,800 longer but the, but the huge benefits of that become apparent 272 00:17:29,900 --> 00:17:31,600 as you're trying to scale these things. 273 00:17:31,600 --> 00:17:35,600 So we deal with it. We deal with the scale and the 274 00:17:35,600 --> 00:17:40,300 volume and the complexity, like any engineering organ. 275 00:17:40,300 --> 00:17:43,100 Is we build tools. We build platforms, and then we 276 00:17:43,200 --> 00:17:45,900 make sure that those scale to those problems. 277 00:17:46,900 --> 00:17:52,100 And right and so now when we do data Goldman Sachs it's like we 278 00:17:52,100 --> 00:17:56,000 do it on the platform and everyone knows like okay now the 279 00:17:56,000 --> 00:17:59,000 quants are going to get their data you know and days instead 280 00:17:59,000 --> 00:18:03,100 of months because we have set up the right platform constructs 281 00:18:03,100 --> 00:18:05,900 and Engineering constructs for that. 282 00:18:05,900 --> 00:18:11,300 So again it's not a perfect Silver Bullet, answer of like as 283 00:18:11,300 --> 00:18:14,900 things grow like, you know, there's some there's some 284 00:18:14,900 --> 00:18:18,500 equation but we have Seen that the investment in the platform 285 00:18:19,500 --> 00:18:22,500 and the tools and the workflows. That's that's the thing that 286 00:18:22,500 --> 00:18:26,400 helps us scale again. I find it really fascinating 287 00:18:26,400 --> 00:18:30,500 that the platform and as you said, the, the workflows play 288 00:18:30,900 --> 00:18:37,300 such a crucial role, but I suppose it's entirely logical 289 00:18:37,500 --> 00:18:41,900 that when you need that data when that data is so important 290 00:18:41,900 --> 00:18:44,900 and it needs to be right? And it needs to be consistently, 291 00:18:44,900 --> 00:18:48,100 write that Need that infrastructure and the 292 00:18:48,100 --> 00:18:51,900 automation to make it happen at that, at that level of quality. 293 00:18:52,400 --> 00:18:55,400 That's exactly right. This is my personal view coming 294 00:18:55,400 --> 00:18:59,200 from an engineering background. I solve pry eyes off, scale, 295 00:18:59,200 --> 00:19:02,600 problems of all your problems and complexity Problems by 296 00:19:02,600 --> 00:19:04,900 breaking it down into engineering steps and 297 00:19:04,900 --> 00:19:07,700 Engineering platforms. And so that's that's sort of 298 00:19:07,700 --> 00:19:11,600 been our ethos of this whole legend platform is really about 299 00:19:11,600 --> 00:19:13,800 that. How to attack that problem from 300 00:19:13,800 --> 00:19:16,500 an engineering lens. We have another Another 301 00:19:16,500 --> 00:19:21,600 excellent question from Twitter. This is from Lizabeth Shaw, who 302 00:19:21,600 --> 00:19:26,100 is now alluding to governance which you mentioned earlier, but 303 00:19:26,100 --> 00:19:27,900 we definitely should talk about that. 304 00:19:28,100 --> 00:19:32,700 And she says, how do you build in the mechanisms that support 305 00:19:32,700 --> 00:19:36,100 data policies? The point of the platform isn't 306 00:19:36,100 --> 00:19:38,000 that? It just magically makes these 307 00:19:38,000 --> 00:19:41,000 things go away. It's that it makes you think of 308 00:19:41,000 --> 00:19:44,300 them upfront, it makes you think of these Concepts. 309 00:19:45,100 --> 00:19:51,000 As you're doing your data design and your data work and so really 310 00:19:51,000 --> 00:19:54,100 when you when you're on the legend platform, thinking about 311 00:19:54,100 --> 00:19:56,800 your data, the first thing you think is what is your data 312 00:19:56,800 --> 00:20:01,400 contract, right? That's actually the first bit of 313 00:20:01,400 --> 00:20:04,600 the workflow is describe your data, describe how you want to 314 00:20:04,600 --> 00:20:07,300 publish it, describe how you want other people to see it or 315 00:20:07,300 --> 00:20:12,900 consume it how you want the track things like lineage. 316 00:20:14,800 --> 00:20:17,800 All of that is sort of built in baked in, as a first class 317 00:20:17,800 --> 00:20:19,700 concept. So again, I don't want to 318 00:20:19,700 --> 00:20:21,700 oversell the thing. You still have to think about 319 00:20:21,700 --> 00:20:24,200 it, you still have to do it, but the point is, bringing it up 320 00:20:24,200 --> 00:20:29,000 front, instead of hiding it as some secondary thing. 321 00:20:29,000 --> 00:20:32,300 And so, have to think about the entitlements, the security, the 322 00:20:32,300 --> 00:20:35,400 encryption, all of that. You think of, as sort of the 323 00:20:35,400 --> 00:20:38,900 first class assets, as you're just a first-class Concepts as 324 00:20:38,900 --> 00:20:42,600 you're designing, your data flow or your data workflow, or your 325 00:20:42,600 --> 00:20:47,700 data, Production or data consumption patterns, and all 326 00:20:47,700 --> 00:20:51,200 that sort of comes together in in the platform. 327 00:20:51,500 --> 00:20:55,900 It's fascinating again to me because it seems that your 328 00:20:56,300 --> 00:21:02,600 emphasis is definitely placed higher on that infrastructure 329 00:21:02,600 --> 00:21:06,900 and platforms. Then other Chief data officer, 330 00:21:06,900 --> 00:21:13,100 others that I've spoken with, but at the same time, you were 331 00:21:13,300 --> 00:21:16,800 Dealing with a level of data complexity combined with the 332 00:21:16,800 --> 00:21:21,500 speed and the and the financial consequences associated with it, 333 00:21:21,500 --> 00:21:25,600 that I think few other companies would have that set of 334 00:21:26,300 --> 00:21:30,200 combination of circumstances. Exactly like the quality the 335 00:21:30,200 --> 00:21:33,200 data has to be right. It has to be consistently, 336 00:21:33,200 --> 00:21:39,200 right, as you said, and the ramifications of that are pretty 337 00:21:39,300 --> 00:21:42,200 or pretty big. We have another question 338 00:21:42,200 --> 00:21:46,700 relating to to governance. And this is from again, from 339 00:21:46,800 --> 00:21:51,100 Arsenal and con. And he says, how do you decide 340 00:21:51,100 --> 00:21:55,000 what data is good to use? What is not? 341 00:21:55,000 --> 00:21:58,900 How do you address bias? He's in the data you're 342 00:21:58,900 --> 00:22:03,500 collecting and using. And this is interesting, he 343 00:22:03,500 --> 00:22:09,400 says, do the business lines, agree with your conclusions. 344 00:22:10,000 --> 00:22:12,800 So how do you also get everybody on the same page around this 345 00:22:12,800 --> 00:22:15,900 stuff? If we are not some like isolated 346 00:22:15,900 --> 00:22:19,900 team in the corner with like pointy hats, right? 347 00:22:19,900 --> 00:22:22,600 Like doing this in isolation. I think that's the first most 348 00:22:22,600 --> 00:22:26,900 important thing to get. Like, we are hands on keyboard 349 00:22:27,200 --> 00:22:31,200 together with the businesses, making sure that the data were 350 00:22:31,200 --> 00:22:35,100 using first is right, right? But the right for them and write 351 00:22:35,100 --> 00:22:40,100 for the use case, And really is solving their problem. 352 00:22:40,100 --> 00:22:43,900 So, the first point I would make is, you know, aligning ourselves 353 00:22:43,900 --> 00:22:47,400 with business outcomes and with the businesses is the first 354 00:22:47,400 --> 00:22:49,000 thing. So like not being in some 355 00:22:49,000 --> 00:22:53,200 isolated back room, like oh, we know best about everything, 356 00:22:53,500 --> 00:22:57,400 every piece of data. And again, the the financial 357 00:22:57,600 --> 00:23:01,600 domain is so complex that we would never even pretend to do 358 00:23:01,600 --> 00:23:04,100 that, right? You know, we have people who are 359 00:23:04,600 --> 00:23:07,200 highly skilled and finance, Financial areas. 360 00:23:07,200 --> 00:23:11,700 So that we could make those decisions with our our 361 00:23:11,700 --> 00:23:19,200 businesses, like, in sort of a joint venture fashion, but we 362 00:23:19,200 --> 00:23:21,200 would never say we're the experts in everything. 363 00:23:21,200 --> 00:23:24,500 So I think that's the first piece is like, you have to be 364 00:23:24,500 --> 00:23:26,700 connected to the business and the business outcomes. 365 00:23:26,700 --> 00:23:32,000 So that's that's the first piece then, then again the quality and 366 00:23:32,000 --> 00:23:36,400 the governance, right is then becomes an aligning incentives. 367 00:23:36,600 --> 00:23:38,200 Of a joint sort of venture, right? 368 00:23:38,200 --> 00:23:40,400 It's becomes there's a healthy tension, right? 369 00:23:40,500 --> 00:23:44,000 Of course we want to do things at scale and they want to sort 370 00:23:44,000 --> 00:23:45,800 of solve the problem immediately. 371 00:23:47,700 --> 00:23:50,600 But the point is again that, you know, bringing these two teams 372 00:23:50,600 --> 00:23:54,700 together really helps accelerate accelerate that and then get to 373 00:23:54,700 --> 00:23:56,400 the right answers and the right data. 374 00:23:56,700 --> 00:24:01,100 So this then leads to the question about the composition 375 00:24:01,500 --> 00:24:06,500 of your team. It's obvious that you have very 376 00:24:06,600 --> 00:24:12,800 three deep technology and data expertise, but as you just 377 00:24:12,800 --> 00:24:18,200 alluded in order to do your job. Again, I have to assume that you 378 00:24:18,200 --> 00:24:24,600 need equivalent Financial depth of expertise especially when you 379 00:24:24,600 --> 00:24:27,500 go into Concepts like derivatives, as you were, as you 380 00:24:27,500 --> 00:24:32,400 were describing earlier, we do have a team of stress or quants 381 00:24:32,400 --> 00:24:38,900 in the maybe more financial world, who Come from being on 382 00:24:38,900 --> 00:24:42,100 the desk. Understanding how data is used 383 00:24:42,100 --> 00:24:47,300 on the desks and have sort of a stem background, whether its 384 00:24:47,300 --> 00:24:51,900 technology math, physics, whatever who are equivalent, 385 00:24:51,900 --> 00:24:54,700 counterparties in Mighty counterparts, in my team that 386 00:24:54,700 --> 00:24:59,700 work with these groups and actually understand the 387 00:24:59,700 --> 00:25:02,800 financial domain, but cross over to the technical domain. 388 00:25:02,800 --> 00:25:06,000 So that's that's what we call our data design and curation 389 00:25:06,000 --> 00:25:08,500 team. I'm pretty pretty cool, 390 00:25:08,500 --> 00:25:12,500 high-powered, high-powered, crew, that that really is bound 391 00:25:12,500 --> 00:25:16,600 to those business. Was those folks which team are 392 00:25:16,600 --> 00:25:18,200 they part of? Are they part of your 393 00:25:18,200 --> 00:25:23,600 organization or part of the finance or trading organization? 394 00:25:23,600 --> 00:25:26,500 Where do they actually fit in our team? 395 00:25:26,500 --> 00:25:29,200 We have a subset of that team, right? 396 00:25:29,200 --> 00:25:34,400 But those, those sort of people are also all embedded in the 397 00:25:34,400 --> 00:25:37,400 business as well. So there's Bedded what we call 398 00:25:37,400 --> 00:25:41,000 embedded, death strats or embedded quants in various 399 00:25:41,000 --> 00:25:45,900 businesses. But then I have a sort of small 400 00:25:45,900 --> 00:25:51,100 selection of that team that works specifically on data with 401 00:25:51,100 --> 00:25:53,400 those other business teams, if that makes sense. 402 00:25:53,600 --> 00:25:59,200 What are the elements that comprise a successful data team 403 00:25:59,200 --> 00:26:02,800 such as you have? It really follows sort of the 404 00:26:02,800 --> 00:26:06,100 structure of the team like the platform team is high-powered 405 00:26:06,100 --> 00:26:07,900 soft. For engineers right there. 406 00:26:08,200 --> 00:26:10,000 They come from a software background. 407 00:26:10,000 --> 00:26:13,500 Their goal is to make the software bulletproof and build 408 00:26:13,500 --> 00:26:15,700 the right workflows for data engineers. 409 00:26:16,500 --> 00:26:19,600 And the second bucket is basically data, Engineers, 410 00:26:19,600 --> 00:26:23,400 people who are content experts but can use the platform 411 00:26:23,400 --> 00:26:26,100 configure. The platform build data 412 00:26:26,100 --> 00:26:28,800 pipelines, build curation pipelines. 413 00:26:30,600 --> 00:26:36,100 Build data models that then get shared out then the third 414 00:26:36,100 --> 00:26:39,900 bucket. Is really this sort of hybrid 415 00:26:40,000 --> 00:26:41,900 hybrid. Again, I'll use the word strap 416 00:26:41,900 --> 00:26:43,500 because that's what we use internally. 417 00:26:43,500 --> 00:26:48,600 This hybrid team, that's really straddles sort of deep finance 418 00:26:48,600 --> 00:26:53,000 and deep Tech together. And then the fourth is sort of 419 00:26:53,000 --> 00:26:56,700 our governance, you know, framework team. 420 00:26:56,700 --> 00:27:00,300 They're the people who sort of set the policies set the 421 00:27:00,300 --> 00:27:05,600 framework on and and and work with the divisions to sort of, 422 00:27:05,600 --> 00:27:08,100 make sure there. They're working in the bounds of 423 00:27:08,100 --> 00:27:11,600 our framework. One of the topics that we have, 424 00:27:11,600 --> 00:27:17,400 not really touched upon is the notion of data interoperability 425 00:27:18,500 --> 00:27:20,700 and I know that's important to you. 426 00:27:20,700 --> 00:27:24,600 So can you tell us about that? We touched a little bit of pain. 427 00:27:24,600 --> 00:27:28,300 It, that's sort of why a big big reason we open sourced. 428 00:27:28,300 --> 00:27:33,400 Our Legend platform was exactly for that was because we felt 429 00:27:33,400 --> 00:27:38,300 that like if we could bring some Platform standardization in the 430 00:27:38,300 --> 00:27:44,200 industry for us or clients are counterparties that then at 431 00:27:44,200 --> 00:27:48,100 least the discussions about how data works or how it should be 432 00:27:48,100 --> 00:27:52,200 connected or how it should be described in the industry that 433 00:27:52,200 --> 00:27:55,100 we could help push that forward and we can help push that 434 00:27:55,100 --> 00:28:00,000 forward with standards with other bodies, but they could all 435 00:28:00,000 --> 00:28:04,700 do that work in one sort of saying way in in our platform. 436 00:28:04,700 --> 00:28:08,400 Well, not it's not even our For me, in the open source Community 437 00:28:08,400 --> 00:28:13,500 platform and and that will benefit everybody. 438 00:28:13,500 --> 00:28:18,200 So it's a bit, it's absolutely a big, a big, big play, I'll 439 00:28:18,200 --> 00:28:22,000 mention, you know, one project we're doing in the in US 440 00:28:22,000 --> 00:28:27,500 Community which is the open source nonprofit that we gave 441 00:28:27,500 --> 00:28:30,300 gave Legend to. We brought is though, which is 442 00:28:30,300 --> 00:28:33,500 this the derivative standards d'oeuvres body? 443 00:28:34,600 --> 00:28:38,700 They have built a The model called CDM, common domain model 444 00:28:38,700 --> 00:28:44,300 for derivatives and now that sort of that data model is now 445 00:28:44,300 --> 00:28:47,300 available in legend for people to collaborate on in a 446 00:28:47,308 --> 00:28:50,400 collaborative environment out in the wild it's has nothing to do 447 00:28:50,400 --> 00:28:52,400 with Goldman Sachs they're just using the platform. 448 00:28:52,400 --> 00:28:57,600 So these are like pretty cool things that you know we're 449 00:28:57,600 --> 00:29:01,700 trying to do to push the interop and and standardization out how 450 00:29:01,700 --> 00:29:05,700 has the uptake bin in the broader community? 451 00:29:06,200 --> 00:29:11,700 Of of this platform and data. Interoperability the cool part 452 00:29:11,700 --> 00:29:15,400 is that our code is out there we have clients, actually deploying 453 00:29:15,400 --> 00:29:17,500 it using it. We have other counterparties 454 00:29:17,500 --> 00:29:21,900 also looking at it, doing pocs, in their own environment, but 455 00:29:21,900 --> 00:29:24,900 the flip side of that is, you know, building an open source 456 00:29:24,900 --> 00:29:30,300 Community is very difficult. I mean, we vastly underestimated 457 00:29:30,400 --> 00:29:35,700 how much time and effort and energy goes into, you know, 458 00:29:35,700 --> 00:29:38,100 making Making sure people know how to use the thing, the 459 00:29:38,100 --> 00:29:40,700 documentation. Do they understand the value? 460 00:29:40,700 --> 00:29:45,600 Can they contribute code back? Like, all of these things, we 461 00:29:45,600 --> 00:29:49,100 knew it would take time, but I think we sort of grossly 462 00:29:49,100 --> 00:29:53,300 underestimated how much time to build sort of that, at that big 463 00:29:53,300 --> 00:29:57,000 Community around these. These projects, what are the 464 00:29:57,000 --> 00:30:01,800 benefits that you have presented to the community as to why they 465 00:30:01,800 --> 00:30:07,000 should engage in this, it has helped us Date of better at 466 00:30:07,000 --> 00:30:10,100 Goldman, is helped us organize our data. 467 00:30:10,100 --> 00:30:13,500 It has helped break down silos. It has helped the data quality 468 00:30:13,500 --> 00:30:15,600 massively. It has helped data governance 469 00:30:15,600 --> 00:30:20,700 aspects, massively internally. And so we point to sort of those 470 00:30:20,700 --> 00:30:24,600 success stories about how we do it internally that we could also 471 00:30:24,600 --> 00:30:30,500 help clients do the same thing. And you know if there's appetite 472 00:30:30,600 --> 00:30:33,600 of course that we could also help the broader industry break 473 00:30:33,600 --> 00:30:36,800 down those silos as well. We have Another couple of 474 00:30:36,800 --> 00:30:39,900 questions that are coming in from Twitter. 475 00:30:39,900 --> 00:30:47,700 Do you think of your data or your platform as a product that 476 00:30:47,700 --> 00:30:50,600 you can sell, given that you've built all of this 477 00:30:50,600 --> 00:30:52,900 infrastructure? I'll say that's not the 478 00:30:52,900 --> 00:30:57,100 intention of doing it. I think, though, again, we have 479 00:30:57,100 --> 00:31:01,900 seen that as clients use it as other Banks, use it, that people 480 00:31:01,900 --> 00:31:06,100 sort of come back to us and ask, okay, well, what could we What 481 00:31:06,100 --> 00:31:10,600 could the support model B, or, hey, can you host this as a SAS 482 00:31:10,600 --> 00:31:12,300 for us? So that we don't have to deal 483 00:31:12,300 --> 00:31:14,300 with setting up the infrastructure, running the 484 00:31:14,300 --> 00:31:15,600 infrastructure or things like that. 485 00:31:15,600 --> 00:31:19,400 So I think we're still again, just not to oversell anything. 486 00:31:19,400 --> 00:31:23,800 I think we're still in a very crawl stages of even thinking 487 00:31:23,800 --> 00:31:26,800 about that community and how we could help do better and better 488 00:31:26,800 --> 00:31:32,900 but Our intention isn't like Like make money off of this 489 00:31:32,900 --> 00:31:34,100 thing. It's really about the 490 00:31:34,100 --> 00:31:38,100 standardization parts now if people see a large adoption and 491 00:31:38,100 --> 00:31:42,500 see a value, I think it could be an exciting potential 492 00:31:42,500 --> 00:31:46,400 opportunity. We have another question from 493 00:31:46,400 --> 00:31:50,600 Twitter. Can you give any examples where 494 00:31:50,700 --> 00:31:57,100 using the platform has resulted in innovative ways of for users 495 00:31:57,100 --> 00:32:02,600 to use and or combined? the data we sort of took again, 496 00:32:02,600 --> 00:32:08,400 a data-driven approach to helping our firm and our Risk 497 00:32:08,400 --> 00:32:11,700 Managers and our sales. People in our Traders, 498 00:32:11,700 --> 00:32:16,000 understand that risk from a data-driven perspective, right? 499 00:32:16,000 --> 00:32:21,900 We have combined all that data and information into our 500 00:32:21,900 --> 00:32:25,700 platform and have basically given I hate to use this word 501 00:32:25,700 --> 00:32:30,600 but democratized the access to that our team has linked a bunch 502 00:32:30,600 --> 00:32:34,900 of that data together done. Some interesting analysis but 503 00:32:34,900 --> 00:32:38,300 more importantly, we have In that those based sets of data, 504 00:32:38,300 --> 00:32:40,900 those relationships, the relationships between those data 505 00:32:40,900 --> 00:32:44,900 that may not be obvious. To our users hands. 506 00:32:44,900 --> 00:32:49,200 That now, they have the access to that information and now they 507 00:32:49,200 --> 00:32:52,500 could come up with creative things or risk, mitigating 508 00:32:52,500 --> 00:32:57,900 things to on top of that data. So I think that's been, you 509 00:32:57,900 --> 00:33:02,900 know, maybe one that resonates pretty recently here, it's kind 510 00:33:02,900 --> 00:33:09,700 of the reverse of Shadow it that cios didn't like years ago. 511 00:33:09,708 --> 00:33:13,900 Some going back, maybe, five years, you're Ali, taking the 512 00:33:13,900 --> 00:33:18,400 data and I assume the tools and placing it in the hands of 513 00:33:18,900 --> 00:33:21,600 users, so that they can be creative. 514 00:33:21,600 --> 00:33:24,400 They can innovate with with that data. 515 00:33:24,700 --> 00:33:28,000 Exactly. But in a fully governed quality 516 00:33:28,400 --> 00:33:33,400 way, right, where that shadow, it problem isn't really a 517 00:33:33,400 --> 00:33:35,300 problem. We have put all the right guard 518 00:33:35,300 --> 00:33:38,900 rails and all the right governance around that. 519 00:33:38,900 --> 00:33:43,000 So now people can actually innovate in a safe space. 520 00:33:43,200 --> 00:33:47,000 In a safe way but it's not just lock to our team. 521 00:33:47,400 --> 00:33:52,700 Can you give us some insight into what kind of governance or 522 00:33:52,700 --> 00:33:58,600 how do you balance the need for security for privacy against 523 00:33:58,600 --> 00:34:02,100 making that data, available accessible and easy to use for 524 00:34:02,100 --> 00:34:04,900 folks before we do anything with data. 525 00:34:05,500 --> 00:34:08,100 Right? We have first of all very strict 526 00:34:08,100 --> 00:34:11,900 policies and Frameworks about, you know, understanding who's 527 00:34:11,900 --> 00:34:13,400 allowed to see what. Data. 528 00:34:14,500 --> 00:34:17,000 And even if you're allowed to see it, should you see it? 529 00:34:17,000 --> 00:34:18,600 You know, we call it need to know, right? 530 00:34:18,600 --> 00:34:20,400 Like just because you're allowed to see it. 531 00:34:20,400 --> 00:34:23,000 Maybe you don't actually need to know that client information. 532 00:34:23,000 --> 00:34:28,300 So first of all, as an overarching theme, I'd say like 533 00:34:28,300 --> 00:34:32,699 one of the one of the most critical pieces is is you know 534 00:34:32,699 --> 00:34:37,000 writing those policies down and actually then sort of enforcing 535 00:34:37,000 --> 00:34:40,000 them in the platform. But before we do anything like 536 00:34:40,000 --> 00:34:44,199 that is a clear thing and then and we have very clear rules 537 00:34:44,400 --> 00:34:46,600 About who can get access to what? 538 00:34:46,600 --> 00:34:51,900 How, and then again, would be a little broken record encode, 539 00:34:51,900 --> 00:34:55,600 those rules in the platform and make sure that that is like a 540 00:34:55,600 --> 00:34:57,500 Slayer. Like if we get that wrong, 541 00:34:57,600 --> 00:35:01,000 nothing matters, right? So the first thing is to make 542 00:35:01,000 --> 00:35:07,100 sure that that Baseline Works before before giving access to 543 00:35:07,100 --> 00:35:10,500 him. So in other words, as long as 544 00:35:10,500 --> 00:35:17,300 the foundation of governance, A risk control, compliance with 545 00:35:17,300 --> 00:35:19,400 regulations, whatever is necessary. 546 00:35:19,700 --> 00:35:26,300 As long as those elements are in place, then you're able to share 547 00:35:26,300 --> 00:35:30,300 the data. And let people, I was going to 548 00:35:30,308 --> 00:35:33,800 say have free range, obviously that's not the case, but have 549 00:35:33,800 --> 00:35:36,100 enough rain, where they can use it. 550 00:35:36,200 --> 00:35:38,900 Use that data creatively in the service of whatever their 551 00:35:38,900 --> 00:35:42,700 business goals happen to be much better said, then that then 552 00:35:42,700 --> 00:35:45,900 white man eyes that it but Exactly, exactly. 553 00:35:45,900 --> 00:35:51,500 The right mental model. What advice do you have for 554 00:35:51,800 --> 00:35:58,000 folks who are building a data strategy, based on what you've 555 00:35:58,000 --> 00:36:00,900 learned at and done at Goldman Sachs? 556 00:36:01,300 --> 00:36:05,100 Make sure you attach yourself to business outcomes. 557 00:36:06,300 --> 00:36:09,600 The things that people care about at your company, right? 558 00:36:09,600 --> 00:36:14,800 That I think, As a technologist. You know sometimes as 559 00:36:14,800 --> 00:36:17,600 technologists in general maybe I'll make a broad generalization 560 00:36:17,600 --> 00:36:19,700 sometimes. You know, we care a lot more 561 00:36:19,700 --> 00:36:25,500 about the tech than the outcome but specifically for data 562 00:36:25,500 --> 00:36:28,400 strategy, I think it's even more important to over-index on the 563 00:36:28,400 --> 00:36:32,900 outcomes because data sort of becomes this nebulous thing 564 00:36:33,100 --> 00:36:35,800 where now you know like well what does it mean to have a data 565 00:36:35,800 --> 00:36:37,900 strategy? What is that like what is data 566 00:36:37,900 --> 00:36:39,400 even mean? Why do I need that? 567 00:36:39,400 --> 00:36:43,000 Like an abstract? I get why I want information, 568 00:36:43,000 --> 00:36:45,100 but like what are you talking about? 569 00:36:45,100 --> 00:36:48,100 So the first thing, I the first, first thing I talked about 570 00:36:48,100 --> 00:36:52,600 always is attached yourself, the business outcomes and show how 571 00:36:53,000 --> 00:36:58,300 the data and the data strategy actually makes those cheaper 572 00:36:58,300 --> 00:37:02,000 faster, better easier makes money for your clients. 573 00:37:02,700 --> 00:37:06,600 Helps reduce risk, help. Save money, do so. 574 00:37:06,600 --> 00:37:11,500 That's that's the first sort of Baseline advice. 575 00:37:11,500 --> 00:37:17,300 I give to everybody. And then then I think, you know, 576 00:37:17,300 --> 00:37:22,600 but then below that then become sort of okay set out. 577 00:37:22,600 --> 00:37:25,800 What is the platform strategy going to be? 578 00:37:25,800 --> 00:37:28,800 How are you going to actually make the engineering work? 579 00:37:28,800 --> 00:37:33,800 Think about the business workflows and the Pro workflows. 580 00:37:35,100 --> 00:37:37,900 And then, you know, we talked a little bit about the org 581 00:37:37,900 --> 00:37:40,700 structure and the framework. So I think those are also key 582 00:37:40,700 --> 00:37:44,400 pieces, right? Make sure that, you know, if 583 00:37:44,400 --> 00:37:46,800 you're going to do it in an engineering strategy, you have a 584 00:37:46,800 --> 00:37:51,000 strong platform team, if you're going to, you know, have a 585 00:37:51,000 --> 00:37:52,600 Content T. Make sure they actually 586 00:37:52,600 --> 00:37:56,400 understand the domain that you're working with. 587 00:37:59,500 --> 00:38:03,600 And so bringing those like or pieces together. 588 00:38:04,300 --> 00:38:09,300 But, but the number one key is like, make sure you're driving 589 00:38:09,300 --> 00:38:11,600 outcomes. What's the relationship between 590 00:38:11,600 --> 00:38:15,200 your data strategy, and Cloud, where does cloud fit into all of 591 00:38:15,207 --> 00:38:16,700 this? I can't believe we haven't 592 00:38:16,700 --> 00:38:19,700 spoken about that. I think of the cloud again, as a 593 00:38:19,700 --> 00:38:24,900 tool in the toolbox, right? It's not a means to an end it's 594 00:38:25,300 --> 00:38:28,800 in my Arsenal how I want to build my platform. 595 00:38:28,800 --> 00:38:31,100 Yeah. I want infinite scalability, I 596 00:38:31,100 --> 00:38:35,100 wanted it, I want, you know, other people have done, the hard 597 00:38:35,100 --> 00:38:38,800 work on the infrastructure front, I want great databases 598 00:38:39,100 --> 00:38:41,900 that people have built to not rebuild that on my own. 599 00:38:42,200 --> 00:38:46,700 And so to me, the cloud is a great enabler, it's a great 600 00:38:46,700 --> 00:38:51,000 tool, my toolbox, it's a great way to get the scale. 601 00:38:51,600 --> 00:38:56,200 So it's definitely a big part of the overall data strategy. 602 00:38:56,200 --> 00:39:01,700 But again, it just to be clear. I just make sure like, It's not 603 00:39:01,700 --> 00:39:05,000 a thing for the sake of a thing, it's like, okay. 604 00:39:05,000 --> 00:39:08,600 These are great components, great pieces of infrastructure 605 00:39:08,600 --> 00:39:12,100 that other people that I can now stand on the shoulder of giants 606 00:39:12,600 --> 00:39:15,300 execute on. And that's the way I think about 607 00:39:15,300 --> 00:39:16,800 it. It was super important part of 608 00:39:16,808 --> 00:39:21,300 the strategy just to be clear. I can see that you are you're 609 00:39:21,300 --> 00:39:25,500 very purpose-driven, you are always coming back to the 610 00:39:25,500 --> 00:39:26,800 reference point. Why are we doing it? 611 00:39:26,800 --> 00:39:28,100 What are we doing at? What are we getting out of it 612 00:39:28,107 --> 00:39:30,000 from them from the business? Outcome standpoint. 613 00:39:30,000 --> 00:39:36,000 That's Very, very, very clear. And finally, one last question, 614 00:39:36,700 --> 00:39:40,700 what is the relationship between data and building business 615 00:39:40,700 --> 00:39:44,000 models economic models Financial models? 616 00:39:44,000 --> 00:39:47,200 How does how to those pieces fit together in technology? 617 00:39:47,200 --> 00:39:51,100 It comes down to algorithms and data and so like those are the 618 00:39:51,100 --> 00:39:54,300 two big inputs, right? Like, you need the data to be 619 00:39:54,300 --> 00:39:56,000 great. You need it to be clean and 620 00:39:56,000 --> 00:39:59,100 needed to be organized. You need to make sure that like 621 00:39:59,200 --> 00:40:00,900 all those pieces are set, it's easy. 622 00:40:01,100 --> 00:40:04,900 Accessible. It's findable to govern and then 623 00:40:05,200 --> 00:40:10,400 that becomes a major input into the Algos, whether you're doing 624 00:40:10,400 --> 00:40:14,400 forecasting or algorithmic trading or helping your client 625 00:40:14,400 --> 00:40:19,400 with something that those pieces just have to like work together 626 00:40:19,700 --> 00:40:23,700 as a as a team to get the job done. 627 00:40:24,200 --> 00:40:26,800 Again, I'm putting words in your mouth, but this is just 628 00:40:26,800 --> 00:40:29,500 summarizing what you've just been talking about is the 629 00:40:29,500 --> 00:40:32,400 linkage to the out. Outcome and being clear about, 630 00:40:32,400 --> 00:40:34,600 what's the data going in? What's the expected result at 631 00:40:34,600 --> 00:40:37,800 the other and and ensuring that the to match up? 632 00:40:38,100 --> 00:40:42,400 Yep, yep, exactly. And with that, unfortunately, 633 00:40:42,400 --> 00:40:45,400 we're out of time Nema. I just want to say a huge. 634 00:40:45,400 --> 00:40:47,800 Thank you for spending time with us. 635 00:40:47,800 --> 00:40:50,800 I really, really appreciate it. Thank you, thanks for the great 636 00:40:50,800 --> 00:40:53,100 questions. Thanks audience, amazing 637 00:40:53,100 --> 00:40:55,700 questions. I really had a really fun time 638 00:40:55,700 --> 00:40:57,700 talking to talking to you Michael. 639 00:40:57,700 --> 00:41:00,000 So thanks for having me, everybody. 640 00:41:00,100 --> 00:41:03,300 Thank you. For watching, I just want to say 641 00:41:03,300 --> 00:41:09,700 a huge thank you to Nema Rafael. He is the head of global data 642 00:41:09,700 --> 00:41:14,800 engineering and the chief data officer of Goldman Sachs. 643 00:41:15,300 --> 00:41:20,000 Now before you go, please subscribe to our YouTube channel 644 00:41:20,000 --> 00:41:23,700 and hit the Subscribe button at the top of our website. 645 00:41:24,600 --> 00:41:26,900 Actually, you know, the Subscribe button has moved to 646 00:41:26,900 --> 00:41:29,700 the bottom of our website, so hit the Subscribe button at the 647 00:41:29,700 --> 00:41:31,700 bottom of our website. So We can send you our 648 00:41:31,700 --> 00:41:34,700 newsletter and keep you up-to-date on our upcoming live 649 00:41:34,700 --> 00:41:36,000 shows. Thanks so much, everybody. 650 00:41:36,300 --> 00:41:38,900 Hope you have a great day and we'll see you next time.