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

10 AI related accounts you should follow on Twitter: 1. Andrej Karpathy — @karpathy 2. Yann LeCun — @ylecun 3. François Chollet — @fchollet 4. Andrew Ng — @AndrewYNg 5. Lilian Weng — @lilianweng 6. Demis Hassabis — @demishassabis 7. Fei-Fei Li — @drfeifei 8. John Carmack — @ID_AA_Carmack 9. Jeremy Howard — @jeremyphoward 10. Gwern Branwen — @gwern follow these 10 before everyone else does. Let me know who I missed?

@aiseomastery @googleaidevs No. I didn't build this - RamP!

@googleaidevs @RamP Ramp building on this makes sense, finance is where reliable agent execution actually matters

Using the new Managed Agents in the Gemini API, @Ramp built their advanced finance agents without touching the backend infrastructure. Learn more ↓

@riskfreemyth @googleaidevs @tryramp Folks tagging me should "try" a little harder to tag the correct one! - RamP!

@googleaidevs @RamP Classic @tryramp / @RamP mixup! 😂😂

Hey @IndiGo6E , huge shout out to your associate Mr. Teethananda at BLR airport. I had a minor issue with my documents. He helped me to get it corrected. He was quick, knowledgeable and very service oriented. He saved me from missing 1st flight of a 10am toe trip. Thanks a ton.

@11x_official No problem. I'm used to getting tagged incorrectly almost everyday and have learnt to enjoy it - RamP!

@RamP Haha you caught us! Our bad :)

Hey @AxisBank @AxisBankSupport my issue 260506008457 is now pending for 20days and still not resolved. I've tried using the chat, call center, backend team (from the email given by call center), visited branch and escalated to Nodal officer. Looks like Axis bank do not care or your software cannot be fixed. Why are you making it so difficult for loyal customers?

@ghumare64 Thanks for tagging me, but I haven't developed anything called Inspect. Perhaps you meant someone else - RamP!

Google's thinking by Addy is the cleanest survey of harness engineering as a discipline I've seen published, and the most important sentence in it is the convergence claim: "If you look at the top coding agents today, they look more like each other than their underlying models do." Three months in a row, this pattern repeats. @WorkOS's Horizon, @Stripe's Minions, @Ramp's Inspect, and now every major coding agent (Claude Code, Cursor, Codex, Aider, Cline) are arriving at the same scaffolding shape from different starting points. @addyosmani enumerates what every harness needs: Prompts and skill files. Tools and MCP servers. Filesystem and git for durable state. Sandboxes for safe execution. Subagents for orchestration. Hooks for enforcement. Observability for traces and cost. Each ships today as its own integration with its own lifecycle. A sandbox provider has one API. An MCP server has another. Subagent frameworks have a third. Observability collectors have a fourth. The harness is the glue holding them together, and most of the engineering effort goes into the glue. The bet behind iii.dev is that all of these are the same primitive: a Worker. A sandbox is a worker. An MCP tool is a worker. A subagent is a worker. A hook is a worker. An observability collector is a worker. Each one a peer process that connects to a registry, registers functions with stable IDs, and subscribes to triggers. Three primitives, closed vocabulary: → Worker: any process that connects (sandbox, agent, MCP server, browser tab, observability collector) → Trigger: what causes a function to run (HTTP, cron, queue, state change, stream, hook event) → Function: named unit of work with a stable ID When the unit collapses, harness engineering stops being glue work. Adding a sandbox becomes a worker connection. Adding a tool becomes a function registration. Adding observability becomes subscribing to traces on the bus that's already there. Three properties drop out that bespoke harnesses struggle to produce: Live discovery. Every connecting worker gets the catalog of every function on every other worker. The harness reflects connected state because the registry is the harness. Live extensibility. The agent can install a new worker mid-task and use it on the next call. The capability graph stays mutable while the agent is still executing. Live observability. One trace across languages, queue handoffs, and the agent-backend boundary, instead of three systems with timestamp correlation. Addy closes with harnesses becoming "more like compilers." The reframe worth making: compilers are static, and the harness needs to be a runtime. Workers connect and disconnect at runtime. Functions register while the system is hot. Compilation freezes the capability graph at build time, which is the wrong tradeoff for agents that need to install capabilities mid-task. @mfpiccolo, our founder at iii, shipped a related piece on the seven core design decisions every harness encodes. Three of them, in his read, are usually answered backwards. Worth reading alongside Addy's survey. Link: x.com/mfpiccolo/stat…

Skyroot has become India’s first space-tech unicorn after securing $60 million in fresh funding from GIC and Sherpalo, lifting its valuation to $1.1 billion bloomberg.com/news/articles/…

@finn_hulse @aidangch Yes yes. Wrong pic though!

@aidangch @RamP ramp? the Angel Investor? the Global R&D Leader - Cloud | Data | Al/ ML | RF |Semicon | EDA | S/w Architecture Itinerant? the Distance Runner?

surely they can afford @ramp now

@AltCryptoGem @hwchase17 @RampLabs I too am - RamP!

@hwchase17 @RampLabs @RamP curious how they handle agent reliability at Ramp's scale. internal tooling like Inspect usually lives or dies on how fast engineers can debug a bad run. going to check this one

Talked to @ramplabs Head of Applied Research Alex Shevchenko on the Max Agency podcast to learn how @Ramp Sheets was built, their internal agent Inspect, and so much more. YouTube: youtube.com/watch?v=trEM9O… Apple: podcasts.apple.com/us/podcast/how… Spotify: open.spotify.com/episode/49NvGQ…

@LangChain @hwchase17 @RampLabs Nice info. You might have tagged the wrong RamP - RamP!

Episode 3 of Max Agency is here! Join @hwchase17 + @Ramp’s Head of Applied Research Alex Shevchenko and go behind the scenes of Ramp Sheets. @RampLabs

@privy_io @asta_li @chapello Thanks for the tag - RamP!

3/ Privy CTO @Asta_Li presented alongside @chapello PM at @Ramp. Together, they demonstrated how digital asset accounts power a full financial experience inside any product — balances, yield, payouts, cards — with stablecoin infrastructure fully abstracted away.

1/ Last week, @Privy_io attended @Stripe Sessions for the first time as a Stripe company. We announced new products, joined the keynote, and spent two days talking with hundreds of builders and enterprises. Here's what we saw on the ground 👇

Stanford’s 2025 LLM course has 9 lectures covering all you need to start from scratch: architecture, training, fine tuning, reasoning, evaluation, and current trends. youtube.com/playlist?list=…

In 2014, Peter Thiel gave a 1-hour class on how to create a monopoly starting from 0. He explained how: · Google became untouchable · PayPal surpassed everyone · Facebook crushed the competition Watch whole video and bookmark for later

Hey @AxisBank @AxisBankSupport My issue is unresolved for 10days (tried call center, useless chatbot, Email etc.,). Then tried escalating to nodal officer Raghunath Nunna "circlenodalofficer.bangalore@axis.bank.in". This email ID taken from your own webpage doesn't exist. You are not even conforming to @RBI regulations. Never thought the service levels have become so pathetic.

Anthropic pays $750,000+ a year for engineers who can build LLM architectures from scratch. Stanford taught the entire thing in 1 hour lecture & released it for free. Bookmark this before it gets buried.

In 2016, a man with no CS degree quit his job to study for a Google interview. He was an English major. A self-taught web developer. A former Korean translator in the US military. He studied 8 to 12 hours a day. For 8 months straight. Algorithms. Data structures. System design. Operating systems. Networking. Every topic Google asks. He tracked every minute of it on GitHub. He called the repo "Google Interview University." Then he applied to Google. Google never called him back. Here's the wildest part: The repo he left behind became one of the most-starred projects on GitHub. Over 343,000 stars. Used by thousands of devs to break into FAANG. He got hired at Amazon as a Software Engineer. His name is John Washam. The repo is now called coding-interview-university. Inside you get: - A multi-month study plan, week by week - Every CS topic Google, Amazon, Meta and Microsoft actually ask - Algorithm patterns with worked examples - System design from zero to senior - Big-O, data structures, trees, graphs, recursion, dynamic programming - Behavioral interview prep - Mock interview drills - Book and lecture recommendations he personally used - Flashcards, video resources, and a coding question practice plan Self-paced. Free. No course. No paywall. No upsell. Just one engineer's 8-month study log, open for anyone who wants to follow it. If you are preparing for a tech interview, this is the most complete free roadmap on the internet. 100% Open Source. (Link in the comments)

In 2014, Stanford professor Matt Abrahams gave a legendary 1-hour lecture called “Think Fast, Talk Smart.” It has 51M+ views for a reason. His frameworks: • Dare to be dull • Greet your anxiety • The paraphrase is your Swiss Army knife 12 lessons on communication:

@gsifthrive @vkhosla @khoslaventures @OpenAI @Breakthrough @TiEGlobal @isbofficial @StanfordBiz @sunmicrosystems Thanks for tagging - RamP!