Operators Ep 23 Transcript
Hi, everyone. My name is Delian and I'm a principal at Founder's Fund, a venture capital Affirm based in San Francisco. This is Operators where I interview non-VC, non-CEO, non-founder operators that make the startup world go around.
Today, I'm interviewing Nik Koblov, head of risk and financial engineering at Ramp. Prior to joining Ramp, Nik spent three years at Affirm working as the head of bank and new markets engineering as well as 18 years at Goldman Sachs before that. I hope you enjoy the show.
Sweet, Nik, excited to have you on. You're going to be the first person on the podcast who's had to let's say, get to know, or at least deal, with another Asparouhov.
That's right. Yeah.
Hopefully, my brother's entertaining.
He is. He's awesome. He's awesome. Pleasure to be here. Thanks for having me.
Cool. So yeah, the way that we normally structure these conversations is before diving into what you're currently working on in the world of Ramp and some of your prior work in startups, I like to always sort of go back in time and talk a little bit about how you sort of started your career in tech, which was you started studying physics and then, studying computer science and then, getting into the world of sort of just startups back in even the mid '90s. And so, I would love to kind of hear what sort of interested you in computer science, why that out of all things in sort of the early '90s. And then, sort of what go you over to K2 Design, that sort of first startup that you joined?
Yeah. It was a wild trip, man. It was also science. I was studying theoretical physics. I was born into a family of scientists. My father was a physicist. My mom was a research chemist. And I always wanted to be a musician, but my parents prohibited me from pursing that. And there I found myself in the university studying physics. And to be honest, I didn't want to be a scientist. I dreamt of music. I skipped on lectures and went to play in a jazz rock band.
But somewhere around my senior year or my junior year, I think it was my junior year, we started doing lots of modeling in turbulence and high rate dynamics. So my school was known, and it's still known in Europe as one of the premier schools of theoretical high rate dynamics. There's lots of big names came out of, like Professor [Dushuney 00:02:39] and Feynman, even in the '30s. So it's a good old school. It predates Soviet times. I think it was founded in 1902.
And the approach typically with Soviet science, is that you sit with a stack of papers like this and you analytically solve partial differential equations till you basically cannot hold your pen anymore. At that point, you take all of this. There's no personal computers anywhere. We're talking about '90s. You go to a research institution somewhere. So my mom worked at a research chemistry lab and she was like, "Yeah, come on over. We have some computers. You can have it." And usually spend all night basically sitting and coding it.
And after one night of debugging, it sort of clicked. I was like, "Oh, my God. This is fascinating. This is what you can do with that." And yeah, so kind of got there through physics. And it was more enjoyable than writing stacks of paper with your pen at that point in time.
So my family moved here in '94. I transferred to Grant Institute at NYU and switched over to computer science. Although I probably should have went into physics quickly finish it, because I just got so many credits in physics. But I was like, "You know what? I had enough of pure science." I actually wanted to do coding at that point.
So went to the job database in college. And there was a startup literally on 11th and Broadway. And Ramp is on 12th and Broadway. So it was one block away that was looking for a founding engineer. And it was basically a traditional designer, a former actor from Hollywood and a Wharton Business grad, three partners, decided to go online. They were a traditional print shop and there was like, "Well, we need some engineers. We have designers, but we need some engineers to bring this online."
And they came over. My English was so bad, they wouldn't let me pick up the phones when they rang. And I was just sitting down, head down. We IPOed eventually. I didn't even know what IPO was. I was like, "What is academics doing there? I don't know." I was just sitting coding and doing my job.
But it was fascinating, kind of do everything from working on accounting team servers and setting up silicon graphics workstations for designers, working with... We had our first hosting service down in Texas and I was like, "Oh." We run stuff on somebody else's computer, pre-dated the cloud. And but yeah, I was just so focused on work, I kind of almost missed IPOs. I was like, "What's happening with the company here?" I was like, "Okay, why is there's all this excitement there?" But yeah. No, it was a wild experience. It was really, really fun.
It sounds like you had some kind of interesting projects that you worked on there, like the two that you at least list on LinkedIn that are kind of cool. You were doing the live web coverage of the IBM Deep Blue versus Kasparov Games and then, also the Toys"R"Us e-commerce website, RIP to Toys"R"Us. But yeah, I guess, I don't know. This sounds cool. Do you have some fun stories from that?
Definitely. So yeah. So K2 Design was specializing on sort of live events in the early day e-commerce. So two super cool projects we got was the Bring in 'da Noise, Bring in 'da Funk. So my most purely fascinating memory was around the opening gala at Urban Plaza, where I was standing with a spark laptop. Literally, this thing was like 20 pounds with a microphone connected into it and interviewing Susan Sarandon when she showed up. I was like, "Oh, my God." There's all these celebrities coming in.
And we basically built what now you call a blog. At the time it was called [Buet&Bork 00:06:38]. People were chatting about the event, "Oh, the show is going to go to Broadway." It's so big. There's all these actors coming in. And so I built a blog. We built a live video feed. It was all over 56K modem speed. It was pretty fascinating all the compression we had to get out of it. We had a lot of help from MIT scientists.
And actually, we built a relationship that went into the next project which was the Kasparov versus Deep Blue. I think it was one of the first, if not the first, live covered events in the history of the internet. For a split second, we had the highest traffic in the world, or so we've been told. MIT servers that hosted the website went down when the match started. We spent the whole night moving all the resources over to the IBM servers. And then, IBM scientists were helping us out to get the website up.
So at the end of the day, there was no load balancer, so we had a page where we had 10 links to 10 servers. So basically, just go to the landing page, like, "Okay, we'll just rely on people randomly clicking on this thing. And then, we'll go in and get to their version of the website on server three. And that will hold through the duration of the games."
So I coded live commentary. So there was IBM speech-to-text software. We streamed in the commentary. We parceled off files from Deep Blue. I was physically sitting in the Philadelphia Convention Center during the game in the same room with the IBM scientists. And there was a backup Deep Blue. So when I walked in, there was this big cabinet in the corner. I was like, "Oh, is this Deep Blue?" And they were, "No, no. This is the backup. The real one is in West Chester in the IBM center."
And then, the scientists, when they got to the end game, Deep Blue, I think it was the second game, it was a notoriously bad at End Game. End Game is hard. And you could see at the end of the match, the scientists literally pacing around like they are playing chess. It's like, no. I think at this point, it's these guys versus Kasparov. And I guess, I know who's going to win. But they're all nervous and the computer is not coming up with the good moves. So that was just eye-opening. It was absolutely fascinating.
The best part of the story is I go back to school, so I was still studying at NYU, and second semester of my network class an IBM dude comes in. The professor invites him. And he talks about load balancers on network routers and he said that, "We built this product that I could describe so we could all take notes. It's fascinating architecture. We can now do load balance as part of the network appliance. It's pretty cool, especially for the '90s." And I raised my hand. I was like, "What made you guys build it?" He was like, "Oh, there was this event, Kasparov versus Deep Blue, and the website went down." I was like, "Oh my God." After class, I went to give him a hug. I was like, "That's amazing work on this project."
So it was great. It was great to see stuff that you study and stuff that you work on just all come together. That was fascinating to me. It's not so dry and academic. It's like there's real life. There's real events run through real internet. It was pretty cool.
That is so, so cool. Wild. Yeah, so I guess, after that, you kind of made a transitioning career that stayed maybe somewhat consistent sort of since then, which was you kind of shifted into a world of sort of mixing let's say computer science plus sort of financial services. And that's basically sort of what you've done in every step of your career sort of since then. And so, I guess, yeah, maybe walk me through sort of how you moved over to Alliance Capital and then, Goldman Sachs. And was that an intentional hey, I want to sort of go into this world of finance plus computer science? Or sort of what motivated that shift?
Yeah. Yeah, I'm a little bit ashamed of the answer. I basically just went for the pay. I was graduating and all the businesses come on campus and you get in offers. And someone just gave me 10K more. Back in the day, the engineers were getting like 40K was the going salary. And I got 50 and I was like, "Oh my God. I'm so rich." And I went for this company. Little did I know, I almost ended up hating fintech because of this first job. It was like everything that I didn't want. It was a big company. It was a buy side. It was then acquired by I think, Axa, a French insurance company. And then, you basically cannot sense exactly what you're doing and why you're doing it.
So I ended up doing personal projects that were just much more interesting in the evening at the job. And I just ran away after a year. If this is what finance is like, I never want to do anything with that.
But I was in New York City. Where does a programmer go in the '90s in New York City? So again, more banks reaching out to me. And then, I ran into Goldman guys and my sort of aftertaste at this point was kind of like, I'm going to quote of our Ramp founding engineers, Pablo Meier. He has a great quote. Becoming software engineer because you like to code is like becoming a butcher because you like animals. It is not what you think it's going to be. It's not going to be a glorious. I write the most amazing architecture all the time. I solve incredible problems. It's a lot of grunt work.
So I sort of was getting that first realization. I was like, "Oh, that fun stuff I used to do at K2 Design, they don't do it in finance."
And then, I ran into Goldman guys and I was blown away. It was a small panel. It was a hard problem. I did not solve it actually. I had a bug. They gave me a benefit of like, "Okay, can you point out where your error is?" I was like, "Ooh, you didn't just turn me around." I was like, "Okay, yeah sure. I screwed up right here." I really appreciated the transparency. And the team was small. It was like, "How big is the team?" It was like, "Well, there's two guys leaving so we're hiring you. It's going to be you and your boss and you're going support $20 billion worth of fixed income funding." And I was like, "Whoa. Are you guys really sure that you trust me with this?"
In about a week, I had the head of fixed income technology sitting on the back of my desk because I incorrectly parsed the head of positions. And the whole fixed income trading desk were trading off the wrong inventory for a day. And I was like, "Okay, so this is real world now. This is real problems."
And this whole realization of your love for animals and you're a little bit of a butcher. So you got to balance these things out. And sort of the weight of responsibility of how much ownership Goldman impairs to their engineers was just incredibly sobering and fascinating. We had complete freedom to do all the technical decisions back in the day and I think, up till now. Goldman really operates as a collection of startups in the same building. So I had nine traders, five operations people and my boss and that was Goldman. It wasn't 30,000 people. That was the folks I was talking to on a daily basis. And we decided to try Java. We decided to retire the old X Windows from traders' desktops. And actually went to a talk how to email and started to modernize the environment a bit.
But I think that kind of scary freedom, I think, was just incredibly, incredibly sobering there. And then, little did I know, the reason I stuck there for that many years is just the amount of problems that are never ending. They're just fascinating and usually, just how sharp people are, how sharp financial folks are. I was sitting on the trader desk on the 26th floor in H5 broad, and was sitting between the two REAPA traders who were literally picking up the phone, calling their clients. And they spent the first 10 minutes talking about their family. Then they called the next client, the first 10 minutes talking about their family.
I was like, "Why are you guys doing this? Why are you wasting time? You got to trade." They go, "This is how you trade." I was like, "Really?" Wow, if shit hits the fan, and they haven't done business every day with these 10 customers, they're not going to do business with me when the market turns down. So a lot of relation management, a lot of understanding what the traders need. Having funding traders stand up in the middle of the trading floor and start yelling REAPA quotes across the floor. I was like, "Why are you doing this?" "Well, I got tell the prices to them." I was like, "Well, here's the screen. You can just type into this and they will pop in on the salesperson's desk." It was like, "Oh, my God."
So there was no product managers telling you you have to build a product. You actually have to learn everything for yourself. You need to understand. And it was all internal, right? It was all internal tooling. You built something you can see the trader using it immediately. It was just fascinating. Fascinating still.
So yeah. You stayed at Goldman for 18 years, which is a phenomenally long run and especially I feel like and today, Silicon Valley area, people do not stay at companies for nearly that long. I guess, yeah, I mean I love that you got to go. You got to be there through the sort of original tech crash. You got to be there through the '08 crash. I imagine getting to sort of be in those rooms and creating technologies, people reacting to those situations. There must have been some really awesome stuff in particular around those two crashes. Anything? Lessons learned or things you'd want to share sort of, from that period of time?
Yeah. So I think that's a great question. What I found is engineers come up with the most ingenious stuff during the crisis. Yeah, living through this crisis was the biggest learning. Building technology that amplifies your core strength as the company during these times and kind of competing with the off-the-shelf solutions.
So I remember taking the train with [Anawa Kwantz 00:16:57], who came from Bear Stearns. And I was like, "Hey, so you went from Bear. Bear crashed. You went to JP. You came to Goldman. Why? Why all those moves?" He goes, "Well now, Bear and Goldman are the only two banks that take real risk. It's just boring everywhere else." It's like, wow.
So there's all this pressure in engineering to support that sort of prop technology. Even servicing, like collecting repayments on the loan, you can do it in a way that it's more competitive. But it was pretty fascinating that engineering can unlock kind of so much business potential.
And so, after being at Goldman for obviously an extended period of time, what started to get you interested, or how did this sort of switch over to Affirm happen? What got you sort of interested in jumping ship from Goldman and trying out the world of startups sort of yet again?
So the story goes, I found myself in Gary Vaynerchuk's office. So my best friend worked at VaynerMedia. And I started getting interested in fintech at around 2015. So I mean, it was interesting to see all the sort of alternative sources of liquidity coming up and lending marketplaces coming up.
But from 2009, 2010 to 2015, there's really nothing new that was created. So there was LendingClub. There was Prosper. There was Earnest. There's LendUp. And sort of it was all about lending. I wasn't super excited about that.
And kind of had this strong feeling that I wanted to talk to somebody. So I turned to my friend, Dennis [Hospov 00:18:38], and he worked for Gary. And I was like, "Do you think Gary can help me debug this a little bit?" And he's such an inspirational speaker.
And I found myself in Gary's office. And Gary was like, "Yeah, I know. I hear you. It's kind of all the same sort of blueprint of lending marketplaces. But you should talk to Max." And I was like, "Why?" He was like, "Max [Lechtan 00:19:04], he's doing crazy stuff. He's lending his own money. It's a point of sale. It's a pretty crazy product and they're doing it completely differently."
I got connected with Lieborg, the CTO, who worked with Max at Slide and they found themselves, after Slide acquisition in Google. And Lieborg just started as a CTO. We immediately clicked on the phone. I flew into San Francisco. And I was like, "Oh my God. This is such a breath of fresh air. This is just unbelievable both on the technical side as well as just simplicity. Like a product that has no fees? You have $100 at 15%. You're giving me 115 after a year. That's all there is to it." And you give the product where it needed most.
So also, kind of a guy from the B2B institutional finance, so consumer. Even though obviously, the regulatory side is much more stringent. There's FTB regulations and all of that. But I think it was just fascinating to try. And as an engineer, I was in my early 40s and never worked in Silicon Valley. I was like, "You know what? If I don't try it now, then when?"
But also, having a family, having two kids, I was a little bit risk averse. I kind of want a product that already found market fit that kind of made it so that I don't take that much risk. But I think the stage was just right in 2016 where Affirm came out of sort of wandering, looking for the product fit. And there were Affirm of merchants that were already onboarding. But there was a lot of foundations that were not built in financial technology and this is what my strength was. So we started putting in all the right pillars for the company to scale to where it is now.
And can you talk a bit about why you found the underlying let's say, business model of Affirm, where you're doing that sort of point of sale lending as a more attractive or innovative model versus the lending marketplaces? What gave you more confidence in that type of model let's say, sort of working better or being more differentiated versus just a pure lending marketplace or let's say, some of the other sort of fintechs that were starting to rise up at that time?
That's a great question. I think what was very hard to see from the outside, so I didn't know it coming in, it was just an excitement of very more different lending product that kind of had a purpose. Very different. So Affirm doesn't give cash, right? So actually, you borrow money but you're never in the possession of cash. And it took me about a year working at Affirm to realize how powerful that is because you actually not interested in...
So it's apparently a less risky product because of that. People have allegiance to a brand such as Peloton, right? So you're far less likely to default. But also, the most important thing that it does, it creates a network of incentives that otherwise, does not exist, right? So the merchants want to be on the Affirm network because they getting higher [inaudible 00:22:16]. They're checkout volumes are increasing because of that. It's the best use of marketing dollars. It subsidize your APR. So that why run an ad if you can have Peloton pay out for 39 months or 199 a month for free?
So it was sort of this kind of brilliant combination of incentives that you create in a marketplace, you create in a network where merchants wants to be in because they gain customer. Customer wants to be in because they actually get the best credit product. And investors want to be in because they get in a very well-performing set of receivables and return on their debt.
So my first project was actually, my first Morgan Stanley debt facility, there was a 100 million to fuel the growth. And that one kind of ran at the breakneck speed with some credits, Moore Capital and all the other partners that our firm now has with over 10 facilities, three securitizations and a variety of finance and debt available. So that by itself is proof that this... It was hard to prove. I remember doing our first securitization. We couldn't explain to the rating agency all our revenue streams. We're making money on this. And we're making money on the merchant. And sometimes, the borrower pays interest, but sometimes it's 0%. And it's like, what kind of product? It doesn't fit into your standard square box.
But to me, I mean, that's a true innovation when you scare the rating agency a little bit. It's like, yeah, yeah, you definitely did something right.
And then, can you talk a bit about just the things that sort of translated well for you from Goldman over to Affirm? It sounds like I'm sure that you were sort of familiar with negotiating $100 million credit facilities, but maybe underwriting loans instantaneously online was probably quite different. And so, yeah, maybe talk through that. And in particular, just how did Affirm think about underwriting this, especially when it's such a sort of unique product?
Yeah. So for first part of the question, things from Goldman that were kind of useful, I think the thing that I was surprised most is that engineering cultural-wise, it wasn't that different. It's probably because kind of Goldman uniqueness within the investment banks, technology, that sort of environment where we were just so empowered to make decisions.
So when I joined the firm, it was like, "Yeah sure." My team makes all the decisions about how we bill servicing, how we do merchant integrations. So that was very familiar. I was worried about adjusting to the pace, but the pace was the same. It was right. It was fast. There was less support, obviously, less infrastructure. We made a lot of early infrastructure mistakes, underappreciated scale. That probably was the scariest when you realized that you introduce a new state on the payment and you just blew up your date sense model because you didn't invest in building an API in between the data scientist and the financial engineers. It's super, super scary. And the models started marking their own decision. It's like, "Okay, it's going to decline the customers instead of approving it." It was like, "Oh, I just did it with this merge? Oh my God. How do we write a test around it?"
The underwriting side was... As I was saying, I did run into my good friend, Delian, I think, who ran [inaudible 00:25:59] in Palantir. He is the head of risk, is kind of doing what I'm doing now at Ramp. But the underwriting was fascinating.
So if you look at your traditional sort of line of credit underwriting, it's actually fairly complicated because of the terms and the open nature of the line. Making underwriting way more transactional for a very well-defined amount or a well-defined term, complete absence of fees, tying this particular swipe sort of Affirm in credit card terms, tied to this particular merchant and the unique incentives that merchant provide in this transaction, all right, actually unlocks a lot of flexibility. And that sort of transactional point-in-time underwriting is still pretty powerful where we always consider sort of the merchant risk and the individual credit risk together. And that combination is pretty unique for each purchase.
Nik, can you walk through? I feel like one of the things that most people don't appreciate is the underwriting of merchant risk. How did you guys sort of think about doing that? And then, also, how did that plug into also the work that you were doing in terms of were there certain let's say, merchants that were easier to secure ties, versus not? And how did that relate to the actual sort of larger sort of lines of credit that you would then have to go integrate with? I know it's obviously discussing S1 and it's relatively well known, but obviously, Peloton was an extremely sort of large merchant for you all. And so, I imagine in some ways, much easier to underwrite, versus a sort of SMB longer tail merchant probably sort of more difficult and maybe couldn't be securitized the same way. How did that tension play out?
So I think ultimately, the solution came through sort of diversifying the lending partners where some of the early-day partners came back later once we were going to gain volume on scale and started offering... They basically trusted us so much that they offered us a facility that they could present. So at the end of the day, if a particular forward flow buyer who also were selling receivables outright will have a further constrained concentration let's say, on I don't know, fashion retailers, right? Then we'll always have an outlet where we could place that and sell it off the balance sheet or at least finance it through our facility.
Ultimately, great merchants bring great risk. And merchants that do after-market auto parts come with their own risk profile. And at the end of the day, I think initially, the focus on the merchant risk was mostly from sort of recognizing the profile of the customer that they bring. But ultimately, we sold it through having unique financing programs for each merchant.
So merchant be looking at the type of customer that they attract at the AOV and how big the purchases are and the overall risk profile. And then, our risk team looking at the overall risk profile of the business, will offer a certain set of financing terms. So with Peloton, you can go as long as 39 months. But initially, we had no facility that we could pledge Peloton to because it was so unusual and Affirm never did more than 24 months terms. But ultimately, through showing the performance of the return customers and through creating the history of vintages, you do prove to the lender that there's yield in this combination of a customer financing program and a merchant.
And that actually, was probably one of the deepest investment into technology to build out. We sort of collaborated between our merchant engineering team and the risk engineering team and the bank engineering team that I used to run. So almost the entire organization came together so that we can actually have the right pricing model on the value that Affirm offers to the merchant. Traditionally, the focus was okay, so what APR? Is this all consumer raised? But probably, kind of around 2017, 2018, the organization came that there's a unique value proposition to price.
And the effect of price on the network effect, right, that I was talking about before of the merchant, the investor and the consumer. And there's a lot of interesting conversation about who should own this piece of code? It's like it came from risk. No, the merchants should.
Well, we modified it on the merchant side. It was like, no, no, it's super important for the bank engineering team, because we service all that stuff. So in reality, it was such a central piece of code that it kind of where all the secret sauce came together and ultimately, realizing and understanding. And that was unique to Affirm, right? It's probably one the firm and early days Clarion and Afterpay of sort of, you have the B2B value proposition and you have a consumer value proposition and there's completely two separate pricing axes. So that was challenging. That was fascinating.
So after three and a half years, or a little bit over that, at Affirm, you then decided to sort of move back to New York and join this very exciting company that I really like obviously, but I'm a shareholder and extremely biased and convinced my little brother to work there too. But yeah, it would be great to hear sort of how you ended up at Ramp. What sort of got you excited about the company? And then, in particular, I'd love to hear you talk about sort of, what about the underlying let's say, business model or financial model, similar to how you saw something spark your eye at Affirm, what sparked your eye at Ramp?
Given that I think for a lot of let's say, unsophisticated outsiders, it's not always super obvious what's so special about Ramp. I obviously, know why it's special. But I'd love to hear you articulate sort of why you think it's special.
Well, excellent question. Initially, as a conclusion I get almost after exploratory recruiting call, it's like, "Oh my God. I didn't realize, that's what you guys actually after." So I'm kind of happy that's a little bit of a hidden secret until you actually talk to one of the Ramplings.
But starting with the first part of your question, so I think the main factor for coming back to New York was a little bit of a home pride. So we always wanted to come back after a few years in the Bay Area. I love the Bay Area. I love the nature. Hop in the car and be in Truckee in two hours if you drive 90 miles an hour at 5:00 in the morning. That was fascinating. And being a very outdoorsy family, we were just in love with the West Coast and the mountains.
But it was time to go home. And the opportunity came up to build out a New York office for Affirm. We were very busy growing the office in nine months from five to 50 people. Now it's over 100. And I was kind of tired after that. So I was a little tired. I decided to take a break. And Karim called me up and was like, "Nik, you got to meet this founder." I was like, "For the first time in 25 years, taking a three-month break. Let's talk in January." He was like, "No, no, no, no. It's time bound. It'll go away. You have to meet it."
I went for lunch with Karim. Karim pulls out a Ramp card. I was like, "You're kidding me. You just started a couple months ago." He was like, "Yeah, this product is working. Let's go back to the office. I'll show you the demo."
I was blown away of how much was built and how thoughtfully the card product and we obviously did virtual cards at Affirm. So we used exactly the same technology as at Ramp, the same processor. And I was very familiar. I knew what it takes to build something like that. And it was live. It was working.
And the two things that stood out to me instantly on the people side, what Karim said just kind of blew my mind in terms of I was like, "What are you after?" He goes, "I want to build a talent brand that years down the road when these people started their own company, is they look back where they all came from." That was Ramp 2019, Ramp 2020. This is that core. This is the absolutely unique set of talent that engineers seek out.
So when you say, "I want be in payments in San Francisco," Strike comes in, comes up instantly. If you say, "I want to be in fintech in Europe," it's like, okay, there's this. There's that. There's this San Francisco company. Move their office here. Maybe there's early day Venmo and then they got bought by PayPal and the engineering culture sort of dissolved a bit.
But even in the way Karim approaches hiring interns, I've noted in the company they hire so many freshman interns, right, so rising sophomores. It was like, "Why are we doing it? It's a lot of work." Right? People barely have any experience. It's like, "What are we doing? It's going to slow our team down."
And Karim's answer to this, "This top 1% that we hire are going to go back to their class. Their classmates are going to see these people spent their summer at Ramp and they all would want to come here and they will know that this is the first person in the class who just had their internship. And very few of us had exciting first time internships." So this sort of word-of-mouth culture carrier brands, so the people who work here, you don't need to sell the company anymore. It's just sufficient to meet the team.
And then, on the product side, what was super exciting is the fact that you don't see fintech companies... So fintech company model is usually, it's their own model and making money, right? There's a particular revenue model. There's the services that they offer. Very few companies exist where you... Affirm is also one of them, where you actually blend the value proposition directly to your consumer to a degree that you're not gamifying with points. We actually want you to spend less.
And it's like, "Eric, how are you going to build a company where you're trying to have businesses spend less money?" He goes, "Oh, we just started this. We're not really after a corporate card." And that's the secret sauce that we tell here in interviews is that on network spend, it's a convenient way to get to the company's cash flow. So you start to understand their spend. We connect to the bank accounts. We understand the cash situation of the company and we can underwrite them better. We don't require any personal guarantees.
We had this story where one of Ramp investors and customers wanted to put an AWS bill on a card, right? So we put 300K on the card. Amazon calls the business in about two days and was like, "Can you move back to bank payments please? Here's $10,000." And so, the customer calls us, apologizes, "So sorry. We tried." The Wall Street business trader's like, "No, no, no. Here it is. We just saved you $10,000. You got a check from us, right?" So this is value proposition that we just brought. And we're going to get that off-network business.
So expanding holistically looking, actually not seeing the companies either in sort of A/P stage of network spend, or on-the-card spend, solving those kind of siloed problems, which I understand why they're doing it. They're very, very complicated problems. I have friends and former Affirmers who operate in a similar space. And they're focusing laser sharp on solving a particular niche.
What's very ambitious about Ramp is it's a holistic approach to understanding the cash flow, the payments, the expenses, the customer payments, the vendor payments, everything that the customer does. As Eric says, we want to be an extension of your finance team. If you sign up to Ramp, you can hire less MBA folks because we do that for you. We're accounting experts and payment experts. We're expense management and invoicing experts. So that is pretty unique and ambitious and hard. But hard problems have been the fun ones.
Can you tell me a bit about how you think you guys cultivate sort of both mentorship and sort of ambitiousness within the junior team? I mean, I love my brother, but in some ways, it's wild to me that he's dropped out of school after sophomore year at Berkeley. And yet, the features that he's working on at Ramp, that he's almost owning entirely end to end, are things that other companies would be owned by two PMs and 15 engineers.
And so, I can't quite tell. I mean, maybe my brother's just a really good engineer. But I'm just like, "It's got to be something about Ramp that's somehow enabling him to be let's say, such a good engineer or to have taught him to be such a great one." And so, I guess, yeah, what do you think enables just the speed and pace that you guys sort of ship products and ship value to your end customers? What about the company really enables that?
Yeah. I think the secret sauce is in ownership and trust. I don't believe in separating, of having product insulate kind of the end consumer and the engineers. That kind of goes back to my early days Goldmans where I was sitting on the trading desk without any PMs telling me what to do. And our head of credit, Srinath Srinivasan, also comes to us from Marcus Goldman, was the head of risk at the Apple Card. We clicked on the interview instantly. I was like, "How do you do [inaudible 00:40:33] on projects?" Because I interviewed everybody in the room. It was like we don't have PM write the spec and then hand it over to engineers. It's like, no, then the engineer doesn't understand what they do.
So one example with Pablo specifically, is we wanted to automate. We just launched as you guys all heard, the $150 million Goldman facility. And we are automating the draw down, the weekly draw down and receivable sale process completely. Pablo was doing it almost entirely by himself. I sat down with our head of capital markets and we literally we wrote maths for Pablo and we said, "Here's how you balance to borrowing base. Here's what your liabilities are. Here's what the assets are. At the end of the day, this needs to be in balance. You're a smart person. You know the math. Go figure it out."
I was kind of one-on-one with him a couple of days ago. He was extremely proud. He was like, "Nik, I wrote some incredible comments about what all these functions are doing." But it's sort of this trust that... I mean, the kind of inherent brilliance in engineers is something that we try to suss out in the interviews. But with Pablo obviously having kind of gone through the internship with seeing how incredibly bright he is and how hard he works.
And then, I'm like, "All right. So why don't you solve this problem entirely on your own? I'm not going to write a spec for you." I wrote something super high level. He went in and just broke it down to the level that, "Okay, you don't need a PM. You don't need Alex and me. you own the space to the degree that when announce at All Hands that we want to hire a capital markets lead," he got very defensive. And he's like, "Nik, I thought I am the capital markets lead."
I was like, "Yeah, yeah, yeah. But if you want to do it, do it. Sure. It's yours. But we have this new product developed, so let's hire somebody to own federal markets longer term and you can go hop around and learn as much stuff as possible."
I think it's the exposure and the support network that we create, especially in the remote environment. We just had almost every engineer go through kind of peer programming set of sessions where they observe and experience engineer code and then send the code base. You move a little slower with the code reviews because you can't help in person, so you communicate through the written comments.
But the extent to which you can do this effectively and you know... I almost call it communicate for code, teaching for code, mentorship for code. It eventually worked. It worked really well. It lend itself very well to the remote environment. And the fact that you can be focused, really be careful with your time here in the day. So the managers that we have, engineering managers, are super sensitive to how much time people are spending in meetings, whether this meeting is needed or not. And as a result, yeah, you get stuff like expense management and three sprints. It's pretty mind blowing this November. We replaced expensive [inaudible 00:44:00] concurrent in only six weeks with three engineers working on it. It's pretty fascinating. I still can't believe it.
Yeah, it's almost wild to comprehend of yeah, especially somebody who's an end user, you go and you're like, "This is actually quite sophisticated software." And you look at the public comps. It's like, yeah, you guys have built the equivalent software. Sure, not every single little edge case. Not everything is perfectly handled. But still, a very, very solid MVP that is the equivalent of what a many-hundred person engineering team has spent a lot of venture capital dollars to build out.
And yeah, I think part of it... I mean, yeah, I can't imagine there's that many startups in the world that are comfortable with a 20-year-old Berkeley dropout moving millions of Goldman's dollars around with code that's entire written by them. Yeah, you guys are probably the only one that are doing that right now. There's probably not anybody else.
Yeah. It's wild. It's wild.
Well, Nik, I really appreciate you taking the time today and telling me about all the various experiences you've had and the great stories. Really, really appreciated it.
It's been super fun, super fun, Delian, as always. Thanks for having me, and until next time.
Thanks for listening, everyone. If you'd like to support the podcast, please sign up for a paid Substack subscription which we use to pay for transcripts, mics and other improvements. If you have any comments or feedback on what kinds of questions I should ask, who should come on the show, or anything else, please do let me know. Have a great rest of your day.