# Ramp 4/10/2026 July 2022. An old friend, Veeral, DMed me about a startup he was working at called Ramp. I knew Ramp from Twitter memes about b2b SaaS and a (farcical) rivalry with Brex; but still only barely. Ramp went on to change the trajectory of my career. I’m in a place now where I’d like to make this arc legible to myself in retrospect. ![[ramp-team-paint-night.jpeg|]] ![[ramp-team-car-selfie.jpeg|]] ![[ramp-team-sailing-nyc.jpg|]] ![[ramp-team-sailing-closeup.jpg|]] ![[ramp-team-bumper-cars.jpg|]] ![[ramp-team-poker-miami.jpg|]] ![[Pasted image 20260307191006.png|]] ![[ramp-team-dinner.jpg|]] ![[IMG_4656.png|]] ![[Pasted image 20260307191311.png|]] ![[ramp-team-selfie-architecture.jpeg|]] ## My time at Ramp ![[veeral-dm-ramp-recruiting.jpeg|DM from an old friend and mentor, Veeral|300]] I still remember the interview process well. I pushed Pavel on what he perceived the long-term ambitions of the company to be. Patrick and I mostly talked about Hacker News and public markets. I could see Miami in his background — he called from the Founders Fund office. The technical challenges were hard enough to feel good about completing; but I also didn’t care enough about the outcome (yet) to let my nerves kick in. I was disillusioned after working in crypto — on a world-changing mission, that I still believe in! — but in a [resource-cursed](https://en.wikipedia.org/wiki/Resource_curse) environment that had lost its discipline. I’d talked to a bunch of crypto teams & investors around the time — wanting badly to double-down, but finding the same reality-distortion field everywhere. I found Ramp refreshing. I *liked* that “corporate cards” were something Twitter made fun of you for working on; knowing that the *people* working on them, were as good as they were. That the scope was — quietly — to upgrade how business is done, and how businesses run, at large. It also seemed like a business that was conspicuously lean and fast in ways that were unusual at peak ZIRP. The discrepancy felt “alpha”-shaped. --- Leaving my previous golden handcuffs behind gave me this faith that I’d leave Ramp if I didn’t like it. I thought *even if* Ramp were good, I’d probably spend two years there, and leave. I stayed far longer than I thought I would. I joined in October 2022, joining a few other engineers on an emerging “Payments” pod: Shreyan (my manager, creator of Ramp’s Payments Platform), Andrew (in charge of hazing me with his intense crypto skepticism), Clara, and later Leina. Pavel was founding tech lead behind Ramp Bill Pay, my skip-level, and only 21 years old. This excited me — clearly this was a place that rewarded slope with great opportunity — but I was also curious how it’d go in practical terms. He turned out to be a great leader — also a high-output/agency IC — and I learned a *lot* working for him: in matters of [gumption](https://x.com/krrishd/status/2030525286949994519?s=46) and speed. ![[slack-pavel-slackmoji.jpg|We’re a big Slackmoji culture|300]] My onboarding task was to add mailed paper checks as a payment method to Bill Pay. Seemed straightforward enough, and I got it “done” — excited that I was *trusted* to! — but I was in for a hard reality check. We were printing, as the check’s memo, the contents of an “invoice number” field rather than a distinct memo field. We also defaulted to Ramp’s office address for the `return_address`. Our punishment: vendors who couldn’t figure out what the check was paying them *for* (they’d need the memo — think credit card companies, tax, etc), and mailing them back to the Ramp office. My specific punishment: guilt for inviting this hell onto my Product Ops counterpart Pravika, who’d joined the same time as me, and was in the NY office triaging hundreds of physical checks while I was supporting remotely. ![[IMG_4594.jpeg|Years later, we all got fun jumbo checks from our partners. Did you know that as long as the right numbers are in the approximately right places, anything can be a check? You can put the right numbers onto a massive birthday cake and successfully e-deposit it.|400]] The sheer *reality* embedded in this problem space — the wrong bug causing physical mayhem — scratched a major itch for me. I wasn’t abstracted away from what we actually did for people. I got addicted to that feeling of responsibility: I’d learned to email customers personally to triage payments edge cases (eg. step-ups), set up our original monitoring, and compulsively stared at everything that could go wrong. Kat (payments PM), Pravika, and I were CC’ed on every single customer payment failure email — and for a while, directly supported them in thread as they needed, using learnings as input into self-service UX. I’d see the names of the business in the tooling, sometimes that of their counterparty; and I felt a sense of responsibility re: making sure these economic participants - patriots! - were able to get settled up. There was so much *meaning* embedded in the work, relative to anything I’d ever worked on. ![[slack-pavel-couscous-gratitude.jpg|Couscous is our internal token of gratitude|400]] Ramp gave me serious responsibility far earlier than anyone else would, and has always been a place that: - under-hires — to a [perfectly pathological](https://www.ribbonfarm.com/2009/10/07/the-gervais-principle-or-the-office-according-to-the-office/#:~:text=Idealized%20organizations%20are%20not%20perfect.%20They%20are%20perfectly%20pathological.) degree — and makes everyone an *owner* of a major function. before AI made it common sense to. - values slope over intercept — to an absurd degree. My skip-level going from intern straight to founding tech-lead of a multi-billion dollar TPV payments product. Teammates with previous lives as eg. Minecraft modders. ^adb135 A few months into my time, Shreyan put in his notice. He took some time to explore, came back to set foundations that’d turn into Ramp Treasury, and then left again to be Elad Gil’s Chief of Staff. Andrew switched to platformize cards, Leina to lead a new platform around Ledger; Pavel, Rodda, Clara, and Patrick left to [start](https://trymoab.com) [companies](https://silnahealth.com) together. I was given the chance to lead Payments at Ramp, and my first direct report. ![[IMG_4606.png|The day I was assigned my first direct report, Patrick sent me his characteristically to-the-point “Definition of a manager”]] Kat, my PM counterpart, saw room for growth re: my understanding of our customers and the business. She took me along to sales calls, customer calls, conversations with our own AP, accounting, strategic, and corporate finance teams — pushed me to spend more time across the org — and was a major teacher (and friend) to me in the process. I learned that Ramp isn’t a place where engineers receive specific instructions from product, to implement. Ramp engineers make product decisions based on their *own* — developing — knowledge of the customer, market, and primitives — in partnership with product. The same way Ramp PMs now *implement* entire features with tools like Ramp Inspect. Everyone gets to do everything. --- Exciting as it was, it was intense that much of the team I’d started with had left the company or switched to lead new teams. Hard to avoid at a company of natural founders and leaders. AI was also ascendant, recruiting heavily, and a lot of my pre-Ramp colleagues had reconvened at a major lab. I still had a lot to learn in Payments, though, and yet to convince myself that I could lead a function of this importance (having never done it before). I wasn’t going to let myself off the hook, yet. Kat and I were tasked with major international expansion: not just being able to *pay out* international counterparties, but being able to make payments *on behalf of international entities and businesses*, originating payment *locally* across Canada, the EU, UK, and beyond. This was the initiative I felt like I had to deliver to the business, before I could even *consider* anything new. Kat, Tristan, and I spent time poring over half-conformant ISO-20022 docs and spreadsheets; on video-off Zoom calls with “implementation” teams at the banks, plugging everything into a prehistoric version of ChatGPT, building with pre-REST tech my parents would be more familiar with. We got our first international collection rails live pretty quickly. The challenges in making Bill Pay & Reimbursements work — where neither sender nor recipient are Ramp, unlike collections — were business ones rather than purely technical. We had to solve currency conversion in light of slower funding rails; FX rates could move against us by the time a payment would be funded. Debit authorization in new geos. Agreements and flows between partner A licensed to acquire/pull, and partner B to payout, etc. I was surprised at how many of these “business problems” were mine to co-design solutions to, as an engineer. But we got it done, and set Ramp in motion for true global expansion. <div style="left: 0; width: 100%; height: 300px; position: relative;"><iframe src="https://w.soundcloud.com/player/?visual=true&url=https%3A%2F%2Fapi.soundcloud.com%2Ftracks%2F1924265513&show_artwork=true" style="top: 0; left: 0; width: 100%; height: 100%; position: absolute; border: 0;" allowfullscreen></iframe></div> ![[drake-canadian-banks-lyrics.jpeg|Dropped right as Anand K. and I - Ramp’s biggest Drake fans - shipped local Canadian reimbursements|400]] Kat left soon after for new adventures. We added Adam and Chapello on product. We also added Osmar, who’d become my team’s go-to for international payments. The pace started to accelerate. ![[ramp-ad-denver-train.jpg|The lite-rail 2 minutes from my apartment.|400]] Most surreal in this period, though, was Ramp’s brand becoming more prominent. I lived in Denver, far away from “tech.” One random day, *every single* billboard around me started to be a Ramp one. Literally down to the bus stops next to my apartment. On the trains passing by my balcony windows. For like a year or more. I’d leave a long day of work to go on a walk, or to get dinner, absolutely surrounded by Ramp. It was a trippy time. The contrast between Ramp on every billboard in a random city like Denver, and the fact that it was a group of [[2025 Year in Review#^3c6189|three]] of us running Payments. Scaling up massive domestic volumes and operations, undertaking more international expansion. Handling our first bank and partner exams, audit. Taking new partners, licenses and flows live. A variety of new financial products for Ramp’s end-customers. My brain was starting to anneal around payments and Ramp. And for all the stress, I loved it. ![[ramp-ad-denver-bus-stop.jpg|Right outside my apartment, where I thought was the middle of nowhere.|400]] --- Some time in late 2024, Stripe bought Bridge. Friends at the company asked me about it (as “ex-crypto” payments guy), and I had some theoretical answers about the opportunity to [[The great irony of stablecoin#^8da528|displace SWIFT]] (which is the single “traditional” rail without a good steel-man — it actually sucks). Vinai — colleague who shipped Ramp’s first internationally-issued cards — Adam and Subham (product), Toshi (partnerships), and I went to the first “A Very Stable Conference” to hear more. Chapello (PM on Payments) and Jon (financial partnerships) went to another stablecoin conference around the same time. We returned pre-aligned on the inevitable: we should kick the tires. ![[IMG_4530.jpeg|A different stablecoin conference I attended a year later with Alex and Arnab. Maybe the coolest venue yet.]] Jon got partnerships together in record time, we got a proof-of-concept for international payments live pretty quickly. As we entered the summer, Chapello dropped me and some of my peers on the Financial Products team (behind our banking product) into a new Slack channel, for a “hackathon project.” The hackathon started the same day an incoming senior engineer on my team - Gian - started - so he joined us too. ![[IMG_3435.jpeg|Gas fees have gotten ridiculously cheap.|500]] ![[IMG_3459 1.jpeg|By choice!|500]] We shipped a winner. And Ramp tends to be pretty serious about doubling-down on these things: we’d earned the mandate to go and ship this for customers more formally, and we booked a week in Boulder to get it alpha-ready (and that we did). ![[IMG_3444.png|]] It quickly turned into a serious initiative — there was a push for my team to allocate full-time resources. We hired the best person we talked to — Alex — in a matter of a few weeks. He was somehow also based in Denver: so Chapello (based in Boulder himself) and I sold him on the opportunity at an in-person dinner. More surreality: I’d never thought a team — my team — would randomly come together in a place like Colorado. A year later — we have meaningful international payments TPV settling (purely under the hood) via stablecoins instead of correspondent banking. We’ve launched a feature-complete version of Ramp usable *entirely* with these assets (and therefore — soon — usable “anywhere”).   I didn’t expect this gig to come [[My internet money fixation - Psychoanalysis#^057609|full circle]] to crypto — delivering novel financial services to *every* kind of business in America, using crypto for practical, useful, *boring* purposes — but we did exactly that. It was especially meaningful in the context of an arc that started with physical paper checks traveling through USPS ([still](https://x.com/0xdaedalus/status/2042661744301215898?s=46) load-bearing). This was a small dream of mine, come true. As my team started to lean more senior: my role started to shift, into organizational support and contributing payments context & guidance more broadly. --- Three and a half years later, I’d become another artifact of Ramp’s devotion to slope over intercept, as someone who basically came in as a kid, and before I knew it, got the chance to lead. ![[IMG_4266.jpeg|How it’s felt|400]] I became more opinionated about engineering culture, hiring, and roadmap in discussion with Matt and Harry (my manager and skip). I got to forward-deploy into projects for our most eccentric customer cohorts — each one eccentric in [borderline](https://x.com/chapello/status/2038704930777911637?s=46) [opposite](https://x.com/tryramp/status/2029565094502506636?s=46) ways — and reconcile them into a coherent product and platform. I worked on more skunkworks projects with friends like Zack and Arnab, done outside of any of our roadmaps, without upfront justification, and sometimes in contrast to intuitions on the business side. There isn’t really red tape. I sat in the Denver office with Alex — and an incredible GTM team, Nigel, Danny, Henry, and others — and would overhear customer problems on their calls, to which I’d get to hawk products and solutions that hadn’t been announced to most of the company yet. Still, purely as an engineer. I started to go to more conferences than are customary for engineers here. Blame crypto. Ramp is a great place to grow up. ## The Ramp secret ^029d07 I was recently asked what I thought made Ramp so AI forward, relative to almost any other company on the planet. Which it is, zero exaggeration. Here’s what I think. Corporate cards, expense management — these are not problems that are intrinsically exciting to most “builder” types by default. The recurring joke in 2022 was that it was a competitive space, with two players who — on the surface — seemed redundant. When I joined, there also wasn’t the benefit of the present-day AI zeitgeist. So why *did* I join? Why did my *peers* join? Ramp’s secret reputation — if you were in the know — was: - **Hyper-lean + slope over intercept:** Come in with energy and capacity for growth, and Eric and Karim will throw you as much rope as you’re willing to catch. - **Comically fast:** Customer has a crazy feature request on Twitter? Engineer spots it, ships within 12 hours of the tweet, gets back to the customer on Twitter right away. Without a single meeting. Before Claude Code, and for that matter, even before *Github Copilot*. ie. by hand. For the love of the game. - **Stepping on toes encouraged:** You don’t need permission to operate in a different functional area than your own, or on a different product/platform. And there isn’t red tape. Engineers do product, design, and ops work. Product, ops, design people ship features and jam on architecture. Everyone does whatever gets the job done for customers *faster* — and people aren’t precious or territorial about it. - **Humble:** Everyone knows our customers are the ones moving the needle — all the way from local churches and farms to the people building datacenters in space and frontier AI models. And that Ramp’s contribution is to help *them* succeed. Strong filter for low-ego: everyone is service-minded. This was true even in early 2022, before the advent of agents that would pour fuel on the fire. Neither I, nor my friends, joined to work on a particular sort of product. We joined to be part of what seemed like a best-in-class software — value — *factory*. ![[IMG_1219.png|Company Kickoff, 2025]] Eric and Karim’s excellence has always been in their intense focus on customers, and their focus on the factory that serves those customers. The Ramp product customers love — one that has undergone, and is again undergoing radical transformation — was always a *byproduct* of those two things. So it’s no wonder that Ramp brought “[Ramp Inspect](https://x.com/zachbruggeman/status/2010728444771074493?s=46)” — rolling your own background agents and harnesses — to the zeitgeist before it was an intuitive choice, inspiring a thousand companies to follow suit. ^9f3029 It’s no wonder that the non-technical side of Ramp is starting to use Inspect and Claude Code at a comparable rate to its engineers. PMs and [designers](https://x.com/chapello/status/2037181762413985942?s=46) are shipping features themselves. Go-to-market, legal, etc are shipping their own purpose-built tooling and automation. Engineers are asking complex questions of the data, and writing product specs. The software is [upgrading itself](https://x.com/ramplabs/status/2036165188899012655?s=46). Ramp was a “factory” company *long before* AI. That’s why many of us joined! It’s no wonder that in the current epoch: it’s building a Gigafactory for business value, before anyone has wrapped their head around what that even means. Ramp is being reborn right now, with the dramatic uplift in AI capability and the resulting set of possibilities that would have been sci-fi even a year ago. There’s so much economic surplus left for Ramp to produce, for orders of magnitude more customers, globally. [Ramp is hiring](https://ramp.com/careers). ![[IMG_4646.jpeg|ramp.com/careers|400]]