RamPrasad "RamP!" Moudgalya

Created a list of around 100 books that had major impact on me over the last two decades. These are also the books that I've re-read several times. buff.ly/2xaaDLd

@benjneb @sugarjammi @swang_co @southpkcommons @modal @CORGI @Genintuition @datadog I'd not be there. Perhaps you meant someone else - RamP!

@sugarjammi @swang_co @southpkcommons @RamP @modal @CORGI @Genintuition @datadog full? In nyc for the day from Chicago!

would personally try to get into these rooms this month in ny btw: 7/14 @southpkcommons x @ramp design demo night (stacked panel with anthropic, hebbia, browserbase, and clay) 7/15 @modal private dinner (iykyk) 7/16 @corgi x @genintuition game night (cute hosts) 7/21 @datadog x @googledeepmind late-night cafe compute (has a barista bar) 7/23 @thecollectiv3e x @elevenlabs x @stanleybystan game night (cool ppl) 7/24 @andruyeung x @rhobusiness founder physical strength training class (y'all could use this one)

@swang_co @southpkcommons @modal @CORGI @Genintuition @datadog Thanks for tagging me - RamP!

@davidgobaud @OpenRouter I can't see anything. Trust me. You perhaps meant someone else - RamP!

How can @Ramp see open weight model usage from spend data when it shows up as @OpenRouter, AWS, Azure, Google Cloud, etc spend right?

@JustinTurk_CEO Partnered with me!? I'm sure you meant someone else - RamP!

🤝 Partnered with @Ramp. ConStrat businesses now get corporate cards, spend controls, approval flows, and vendor payments, averaging 5% savings. New accounts get a $500 bonus. Financial independence, operationalized.

Independence Looks Like This 🇺🇸 Happy 4th of July. 250 years ago, independence meant refusing to wait for permission. For the builders we serve, it still does. You don’t get it handed to you. You earn it: cash flow, bonding capacity, knowing your numbers cold.

@xNoNewNews @RevelioLabs @arakharazian Thanks for tagging me. Perhaps you meant someone else.

Revelio + @Ramp used actual AI vendor spend across 21k US companies, not job-description guesses. The biggest AI spenders had ~10.2% higher employment after adoption. Low-intensity adopters saw no significant gain. Adoption intensity matters. 🔥@RevelioLabs / @arakharazian

@ADINonline @aixventureshq @huggingface @perplexity_ai @wandb @AmbienceAI @airstreetpress @AnthropicAI @poolsideai @wayve_ai @CompoundVC @Astranis @refactorcapital @coinbase Thanks for tagging me - RamP!

seed funds quietly cooking that nobody talks about: @aixventureshq → @huggingface, @perplexity_ai, @weights_biases, @AmbienceAI @airstreetpress → @AnthropicAI, @poolsideai, @wayve_ai @compoundvc → @AnthropicAI, @astranis, @huggingface @refactorcapital → @coinbase, @boomaero, @solugen @notationcapital → @figma (seed), @clipboard_health, @PostHog @constructcap → @hadrianautomate, @veho, @CopiaAg @cantosvc → @atom_computing, @astranis, @VarunaTech @longjourneyvc → @cluely, @PikaLabs, @figure_robot @liquid2vc → @Rippling, @gusto @workliffe → @webflow, @public, @kalshi @bananacapvc → @Ramp, @mercury @2048vc → @airbyte, @Levels @PatronHQ → @rec_room, @theta_labs

@armandlabs @HubSpot @Microsoft365 @QuickBooks @ToastTab @webflow @figma @RingCentral @Google @telegram Appreciate tagging me with all the biggies - RamP!

I was always the business owner trying to get ahead. I had all the tools. @HubSpot, @Microsoft365, @Ramp, @QuickBooks, @ToastTab, @webflow, @figma, @RingCentral, @Google, @StapeIO, @telegram, all of it. But it still felt like I was surrounded by software instead of actually in control of the business. Then I saw a few hype videos about @openclaw and agent workflows on Instagram and figured I’d try it. It changed my life. Now I can run my business from Telegram. Pull any report. Build internal apps. Automate multi-source reporting. Create sales intelligence. Launch landing pages. Build custom HTML emails. Write workflows. Place ads with clean server-side tracking through GTM, Stape, and HubSpot. Connect the tools that were never really talking to each other. The crazy part is not just speed. It’s that every annoying business problem now feels solvable. We do not have to wait on agencies. We do not have to hire a consultant for every workflow. We do not have to accept whatever the SaaS tool gives us. My team can have an idea today and we can ship it tomorrow. I genuinely feel like we can take over the world with this. I have never been more excited to build.

@Paulina_laba @zocomputer @Venmo @PalantirTech Thanks for tagging me with other elites - RamP!. I'm sure you meant someone else 😀

NY Tech Week wrapped up: - hosted an event with @zocomputer and @Venmo co-founders that was such a success that people did not want to leave - got an insane partnership opportunity with a +$300B company - joined +30 events and met so many amazing people from @PalantirTech @Meta @clay @ramp @cursor_ai and more - tried VR experience for the first time - realized that NY pizza is indeed supreme

AI & ML FREE Courses from Stanford: ❯ CS336 - LLM from Scratch ❯ CS221 - Artificial Intelligence ❯ CS229 - Machine Learning ❯ CS230 - Deep Learning ❯ CS234 - Reinforcement Learning ❯ CS224N - NLP with Deep Learning All course links inside:

@cornelliusdbs Joined me!? Perhaps you meant someone else - RamP!

joined @Ramp to work on AI products. ended up designing their merch too. nobody tells you the random skills carry

@TimDodaro2 @gianna_auer_ Sorry I can't help - RamP!

@gianna_auer_ @RamP IPO please

I joined @Ramp a little over a year ago. Today, we're announcing our Series F. The funding is exciting. Getting to hire the people building what's next is even more exciting. And there's still so much left to build. P.S. I'm hiring another design recruiter! If shaping the future of Ramp's design team sounds fun, let's chat.

@gianna_auer_ Sigh - RamP!

@ebarroca Like what? - RamP!

Making up reason for expenses it’s great @ramp. So many problems with this

@ProfeLuke Sigh - RamP!

i know i can go find out how to install the @Ramp CLI, but this sure would be a nice place to put the instructions (this card is NOT clickable)

@ContentIsHot @lplenglish @tryramp Thanks for “trying” to correct the mistake. Enjoying the attention though I get wrongly tagged each day. Sometimes several times.

@lplenglish small mistake - not @ramp :D but @tryramp 😂

YOU WANT TO KNOW WHY $OPAL WILL HIT BILLIONS? 💥 $OPAL key is right in front of you. unlock your exit from the matrix. the only question is whether you take it 🤷‍♂️😉 Let's talk about existing and potential clients and TAM. confirmed client - LPL @lplenglish - world's largest esports league. revenue already flowing. contracts executed. already happened. potential client - Ramp @ramp ($40B valuation, IPO-ready by end of 2026). Josh met their engineers directly. Ramp doesn't take meetings for fun. but let's zoom out. who else needs this data? AI labs - in an arms race for RLHF behavioral training data they can't generate themselves: -OpenAI @OpenAI ($852B) -Anthropic @AnthropicAI ($965B - just raised Series H today) -Google DeepMind @GoogleDeepMind (Alphabet, $2T) -Meta AI @MetaAI (Meta, $1.4T) -xAI @xai ($200B) Robotics - need human motion, reaction and split-second decision data at scale: -Figure AI @figure_robot ($39.5B) -Boston Dynamics @BostonDynamics (Hyundai-owned) -Tesla Optimus @Tesla ($1T) Autonomous vehicles - real-time human reaction data under pressure: -Waymo @Waymo ($126B) -Mobileye @Mobileye (~$10B) Gaming giants - behavioral data on their own players they never had structured access to: -EA @EA ($55B) -Riot Games @riotgames (Tencent-owned) -Epic Games @EpicGames (~$31B) Josh said it best in the interview: "Labs will buy as much data as possible. any extra data incorporated into the corpus is an extra way to win the AGI race." TAM for behavioral AI training data - tens of billions and growing with no ceiling in sight. you do the math - billions. and it's not hopium. DYOR, verify, bro ✊ gl, hf. $OPAL CA: 2PzS5SYYWjUFvzXNFaMmRkpjkxGX6R5v8DnKYtdcpump Target: Billlllllllllllllions

@Techweek_ @fin @stripe I’d not be making it to this event - RamP!

Sometimes no agenda is the best agenda. Grabbing lunch in Back Bay with @Fin, @Stripe, and @Ramp with a small group of founders. #BOSTechWeek

@mikecarbone @WorkOS @cursor_ai @clay @tryprofound OK. I’ll also register - RamP!

Come hang at the @workos office in SoHo June 9th with me and the team! Featuring additional speakers from @cursor_ai @ramp @clay and @tryprofound 🔥 it's going to be a banger, don't miss it

Some of you know me through my art and gardening column. I am also a co-founder of a Bangalore food tech startup. We are starting a pilot. Please retweet this for good karma! About The Pilot RedTile is a single-person food system that slow cooks tasty healthy meals with zero daily planning or cooking hassle. We’re looking for 30 young people in Bangalore to use it for about 6 weeks. You’ll get: •⁠ ⁠A RedTile unit for 6 weeks •⁠ ⁠Food packs to start — Poha, Upma, Kashmiri Pulao & more •⁠ ⁠Full support from our team All we ask: honest, practical feedback on how it fits your routine. Interested? Fill this form out: forms.gle/ZdxD3ymyEo8RrJ… We’ll reach out if it’s a fit! RedTile: redtile.in

Hey @AxisBank @AxisBankSupport my issue (case ID: 260513376716) is unresolved for a month. Even multiple escalations to the PNO hasn't gotten him or his team out of deep slumber. Do not send a standard "check your DM" nor call me. Get this going.

@MarcBaselga Thanks for tagging me - RamP!

In a recent Supra session, Sebastien Goddijn from @Ramp shared something I keep thinking about. Ramp had a bunch of people using AI. On paper, adoption looked good. But adoption alone wasn't creating the value they wanted. The engineers getting real value had done a lot of setup work around the tools. Context files, MCPs, memory, workflows, examples. All the annoying scaffolding that makes the model actually useful. A lot of non-technical folks were in a very different place. They were opening a powerful blank screen and were expected to figure out the whole environment themselves. That's the part I think many companies are underestimating. They buy Claude, ChatGPT, Cursor, whatever else, and then quietly make every employee pay the setup tax. Seb said some people at Ramp knew the tools were powerful but still felt like "it doesn't work for me." Others were spending 20 hours trying to build a custom MCP server for one workflow. That's a pretty bad place to put a non-technical operator. Glass, Ramp's internal AI productivity suite, is interesting because the chat UI is only part of the story. The part I found more useful was the setup layer around it. The pieces I would steal are pretty straightforward. 1/ Connect the systems first Slack, Notion, calendar, email, files, Salesforce, whatever the person actually uses day to day. Otherwise the employee becomes the context courier, copy-pasting from one system to another and hoping the model has enough to be useful. 2/ Build memory from where work already happens A useful memory system includes the stuff around the person, not only their past AI chats. Role, active projects, people they work with, communication preferences, recurring capabilities, all pulled from the systems where work already lives. Seb showed an example where Glass knew who his manager was, how to contact him, what systems to use, and what context mattered. Small thing, but that is the difference between "send Ben a message" working naturally and the model asking five clarifying questions first. 3/ Make good workflows reusable Ramp has this skills repo called the Dojo, where finance curates skills, data curates skills, and people can share what they build. Only the stronger ones get promoted. I like that balance because it keeps experimentation open without pretending every random workflow deserves to become company canon. 4/ Teach people while they're doing the work Seb's plan mode example was painfully relatable. You can tell people "use plan mode" for months and they'll still forget. But if the product notices "this looks like planning" and points them to the mode right then, that's a much better teachable moment than a slide in an onboarding deck. I don't think the lesson is "go build your own Glass." Seb was pretty explicit that he'd rather use a product if one existed that did all this well enough. The useful lesson is more basic. Whatever tool you choose, someone has to own the setup layer.