Fetch expands credit facility, unveils AI tool FAST

A while back in a Meesho interview, I was given an LLD problem: "Design an online polling system where users can create polls, vote, and view results." As I was learning Spring Boot, I thought — why not actually build it?  That’s how CrowdQuery was born 🚀 A platform for open and anonymous conversations where people can ask, share, and engage without hesitation. With CrowdQuery, you can:  💬 Create or join channels around topics (using invite codes) where anyone can post questions anonymously.  🖼️ Post images along with your questions or discussions.  📊 Create polls and participate in them.  🤝 Interact with other posts. I built the backend in Spring Boot, the frontend in Next.js, used Supabase for Postgres + image storage, and deployed on Azure with Docker & Nginx. Checkout CrowdQuery here : https://lnkd.in/gJQrKQpc

Instead of the usual “HackMIT was a blast” post, I’ll say this: You’re looking at about $88M in yearly college tuition. You’re looking at junior software engineers. The kind of entry level jobs Big LLM is effectively killing. Y Combinator was there - they offered an interview if they replicated one of their startups from the 2022 era (that’s their shelf life I guess) Pete went on stage after them and told them we are masters at copying businesses — Maddie D. Reese just a week earlier won a hackathon by using Toolhouse and Lovable to rebuild Google Drive. So these software engineers did the only possible thing: they ditched Claude Code and started using Toolhouse instead. Why did they switch. Toolhouse is simpler — if our customers can build their agentic flows in days, you can win a hackathon in hours. Toolhouse is more affordable — we do a lot of work behind the scenes to optimize the best models for agentic behavior, and we pass the savings down to our users. Toolhouse is customizable — there’s flexibility in literally any part of the experience. Dario Amodei says there’s a 25% chance AI will be extremely bad for the economy. We’re here to bring that chance down to 0%.

#FridayFunda 2803 Reimagine transformation or transform reimagination? In 2014, the call was to reimagine— to dream in digital colors, to sketch futures on an empty canvas. In 2025, the demand is to transform with AI— to turn vision into velocity, to carve meaning from momentum, to let intelligence amplify imagination. Reimagination opens the door, transformation walks us through. And between the two, every age finds its balance. For those who’ve weathered the tides, who’ve reimagined, transformed, adapted— here’s to patience tested by layoffs’ waves, the whooshing currents of change, where titles vanish like sketches in a code refactor. Yet remember—after every prune, a new branch can sprout, and AI’s dance with humans is only in its opening step.

The no-code revolution isn't slowing down, it's evolving. I've been building with Bubble.io for years, and what I'm seeing now is fascinating. We've moved beyond simple prototypes and landing pages. Today's no-code builders are creating enterprise-grade applications that compete directly with traditionally coded solutions. The shift happened when platforms like Bubble introduced advanced database relationships, API integrations, and sophisticated workflow automation. Suddenly, the "ceiling" that limited no-code development started disappearing. I recently worked with a startup that built their entire customer management system in Bubble. Six months later, they're processing thousands of transactions daily with zero performance issues. The founders? No traditional coding background whatsoever. This tells me something important about the future of software development. We're not replacing developers, we're democratising the ability to solve problems with technology. The next wave will be even more interesting as AI-assisted no-code tools emerge. Soon, describing what you want to build will be enough to actually build it. #NoCode #BubbleIO #SoftwareDevelopment

Why Every Full Stack Developer Should Learn API Integration When I started, I thought Full Stack = frontend + backend + database. But then I discovered the real game-changer: APIs. Everyday apps run on APIs 👇 🍔 Swiggy → Maps API + Payment Gateway 🚕 Uber → Maps API + Payment API + Notification API 🔑 Social Logins → Google/Facebook API  APIs help developers: 1️⃣ Save time → Reuse existing services 2️⃣ Scale → Handle large user bases 3️⃣ Build smarter products → Combine services 4️⃣ Stay in demand → Industry loves API skills The future isn’t about building everything from scratch. It’s about connecting the right building blocks. Have you built a project using APIs? Which one challenged you the most? #FullStackDeveloper #DSA #Programming #CareerGrowth #ProblemSolving

A Hidden Gem for Nerds and Curious Minds 🤖 You know those treasures hiding in plain sight that most people just scroll past? This repo is exactly that. It’s a hidden gem packed with real-world wisdom that many miss without realizing how valuable it is. - 300+ case studies 📚 from 80+ leading companies like Netflix, Airbnb, and DoorDash - Each one dives into real ML system design 👨🎨—no fluff, just how they actually built and shipped it What I love about it: - Not just theory 🧮: These are practical, battle-tested examples straight from production - Peek behind the product 🍦: See how ML is embedded in features, workflows, and ops to boost quality and efficiency - Learn by example 💡: Way more engaging than dry concepts—you get living, breathing use cases you can adapt If you build, design, or wrangle ML systems—or just want to see how the best do it—this is your playground. Go treasure hunting 💎 here: https://lnkd.in/dyYSzVqc P.S: image by Qwen #GitHub #Repository #Learning #ML #Example

When Claude Code Joined Our Team in the Middle of a Tech Scavenger Hunt Yesterday, our randomly assembled team of four engineers joined an intense tech scavenger hunt at InMobi Encore2025 - packed with coding challenges and technical riddles. The team: Me (Staff Engineer – Frontend) An SDE2 – ML An SDE3 – Backend Another Staff Engineer – Backend One puzzle stopped us cold: two giant parquet datasets (~750K rows each) hiding a riddle about bids, floors, installs, and a “final pattern” buried deep inside. My teammates? Instantly spinning up Databricks, notebooks, pandas. Me? No relevant setup. No DS tools. Just sitting there, watching the clock tick, feeling like dead weight. 💡 The Lightbulb Moment "Why not try Claude Code?" I dropped the datasets and riddle in with few more lines of instructions. And while we were still stuck on the first step, Claude tore through: Filtering down to iOS/Android Checking bid floors Finding the winning ads (render + click + install) Merging on impression_id Revealing a Fibonacci sequence in time_to_install Answer: 34. “Guys, just enter 34.” Suspicious stares. A pause. Then… Correct! That single AI-powered leap saved us 10-15 minutes and flipped the game. From that point on, Claude became our fifth teammate. Takeaways: In tech, the boldest move is often trying what looks unconventional AI can be a genuine partner in problem-solving Resourcefulness > perfect setup And sometimes, success in uncharted territory comes from being willing to try what others dismiss I may not have written a line of pandas, but being open to AI meant I could unlock value my team didn’t expect. PS: We didn’t win the contest, but we came very close — only to be tripped up in the last round that required us to actually run around the office for the answer. Turns out, AI has its limits 😉 #AI #Engineering #ProblemSolving #Teamwork

At Euka AI, our engineers work directly with customers. No product managers. No requirements documents. No approval layers. Customer reports a bug? Engineer jumps on a call within hours. Feature request from a paying client? We ship it the same week. This sounds chaotic to people from traditional tech companies. But speed trumps process when you're building something new. I was a product manager at ByteDance, DoorDash, and PatPat. I know how that role can add value in mature products with complex stakeholder alignment. But for early-stage companies? Every layer between the builder and the customer slows down learning. We're not building features for millions of users with conflicting needs. We're solving specific problems for brands we are actively partnered with. The feedback loop is direct. The iteration speed is everything.

𝗧𝗵𝗲 𝗳𝗲𝗮𝘁𝘂𝗿𝗲 𝗲𝘃𝗲𝗿𝘆𝗼𝗻𝗲 𝘀𝗮𝗶𝗱 𝘄𝗼𝘂𝗹𝗱 𝗸𝗶𝗹𝗹 𝘂𝘀. There was no customer request. No market survey. Just me, seeing a gap we had to solve. I fought for months. Got 2 devs for 1 sprint. 𝗖𝗼𝗺𝗽𝗮𝗻𝘆 𝗿𝗲𝗷𝗲𝗰𝘁𝗲𝗱 𝗶𝘁. 𝗪𝗲 𝘀𝗵𝗶𝗽𝗽𝗲𝗱 𝗮𝗻𝘆𝘄𝗮𝘆. That “bad idea” became the biggest revenue driver in the company. Here’s the lesson I took: 1. Silence isn’t proof → nobody wanted a phone that’s only a screen either… until they held it. 2. Small bets prove big visions → 2 devs, 1 sprint was enough to show impact. 3. Vision beats validation → sometimes you need to act before the market asks. 💬 When have you shipped something nobody believed in — until it worked?

🚀 Thrilled to back Floot, fresh out of YC S25. Most no-code platforms were bolted onto developer stacks — which means non-coders often get stuck, frustrated, and give up. Floot flipped the script: they rebuilt the runtime from the ground up so AI can write, debug, and ship production-ready apps. The result: anyone — from a lawyer to a doctor to a union organizer — can now build real software, not just toy prototypes. Why we invested: • Timing: AI is finally reliable enough to make app-building accessible to millions of non-technical creators. • Team: Engineers who’ve scaled products at Retool, Asana, AWS, and bootstrapped to millions in ARR before. • Vision: Become the AWS for non-coders — an integrated platform where ideas don’t get stuck. We’re excited to partner with Yujian, Edward, and the Floot team as they open a new chapter in software creation. 👉 floot.com