Meesho's AI Strategy: Continuous Learning and Improvement

In an era where i hear the word 'AI' being abused with more style than substance- i must say it is really refreshing to read this interview of Sanjeev Barnwal. A recommendation engine running 6 trillion inferences a day, GeoLLM, the social imperative of AI and some thoughts on the digital behaviour of the next 100 million digital users..#GoodReads #AI #B2B #Data

In this conversation with FE Futech, Kiran Kumar Katreddi, Head of Platform Engineering and Agentic AI at Meesho, explains why the future belongs to sustainable intelligence systems that continuously learn, optimise costs and deliver measurable impact at scale. Key takeaways: 🔹 BharatMLStack has reduced AI computing costs by 60-70% 🔹 Over 70% of Meesho's code is now AI-generated 🔹 Agentic AI is improving logistics, customer support and commerce outcomes across the platform 🔹 Competitive advantage lies in economically viable AI systems, not just larger models ✍ - Abhishek Raval 🔗Read the full interview to understand how Meesho is building AI for scale, efficiency and business value- https://lnkd.in/gqaubVrg #FEFutech #ArtificialIntelligence #AgenticAI #Meesho #MachineLearning #DigitalCommerce #GenerativeAI #Technology #Innovation #Engineering #AITransformation

🚀 Innovation Powered by AI. Built for the Future. We're excited to showcase the Drytis AI Platform—a modern AI ecosystem designed to help businesses automate workflows, enhance productivity, and accelerate digital transformation. At Drytis, we're building intelligent solutions that simplify complex processes, empower teams with AI-driven insights, and enable organizations to innovate with confidence. 🎥 Watch our latest platform demo to explore how Drytis combines cutting-edge AI with an intuitive user experience to deliver smarter, faster, and more scalable solutions. This is just the beginning of our journey. Stay connected as we continue to introduce new capabilities, product enhancements, and AI innovations that help businesses unlock their full potential. 💬 We'd love to hear your thoughts. Share your feedback in the comments or connect with us to learn more. #Drytis #ArtificialIntelligence #AI #Innovation #DigitalTransformation #Technology #Automation #EnterpriseAI #MachineLearning #FutureOfWork #BusinessTransformation #Software #SaaS #ProductInnovation #TechInnovation #startup #AIisthenewfuture #GlobalAI Amazon Microsoft Flipkart Myntra Netflix

In the AI era, what still deserves human attention? Everyone is talking about AI making teams faster. And it's true. Today, businesses can use AI to: • Generate creatives • Build dashboards • Automate support • Launch products Which means teams can ship more than ever before. But sale and peak seasons create an interesting challenge. When AI helps you do more, what do you prioritize? One pattern I've noticed is that high-performing teams often prioritize the things customers don't see on the surface. ✅ Reliability over feature velocity ✅ Conversion and success rates over cost savings ✅ System readiness over last-minute additions A brilliant campaign means very little if customers have a poor app or website experience. A perfectly optimized funnel means very little if payments fail at the final step. That's why, heading into peak sale periods, many businesses focus on metrics like operational visibility, response times, system resilience, and payment success rates. Take payments, for example. Businesses don't just need a gateway that processes payments. That’s why more and more businesses are investing in payment gateways like Razorpay who are investing not just in reliability and payment success rates, but also in an agentic dashboard that makes visibility easier and conversational for a founder that is already short on time. So founders now know what went well, what needs attention and what action to take next. Because the challenge now isn't doing more. It's knowing what matters most. Check out more details here - https://bit.ly/3SKc1O6

MEESHO 🤝 MOBAVENUE AI As Meesho continued to scale, transparency and agility became key to unlocking new opportunities. With a collaborative approach and a growing suite of AI-powered solutions, Mobavenue AI helped Meesho: • Build with complete transparency • Enable rapid experimentation and faster iterations • Continuously test, learn, and scale together Because the strongest partnerships aren't built on a single solution. They're built on trust, innovation, and a shared appetite to keep experimenting. #MobavenueAI #Meesho #testimonials #Experimentation #scale

This week felt like a small but important plot twist in India’s AI story. Pramaana Labs raises a $27M seed to make AI provably correct in high‑stakes work like law, drug discovery, and tax. MoEngage buys Aampe to pull a real 1:1 AI personalization engine into its platform and win migrations from heavyweights like Salesforce and Adobe. Different deals. Same signal: Capital is rewarding reliability, distribution and measurable ROI. Not the shiniest demo. If I were pitching an AI startup to Indian investors right now, I’d over-index on three things: 1) Metrics - Hard ROI: revenue lift, cost per outcome, payback period - Product health: retention cohorts, expansion, NRR - Unit economics: gross margin per use, infra cost per decision - Reliability: failure rate, latency SLOs, eval scores on real tasks 2) Narrative - Why now: specific regulatory, infra or customer shifts - Wedge: 1 narrow, painful workflow you dominate - Why India: access to talent + data + price-sensitive customers that force discipline, then go global from there 3) Governance - Clear risk register for your use cases - Audit trails of model decisions - Red-team / eval reports, not just marketing claims - Compliance-ready data flows and retention policies What founders should stop saying in decks: - “We’ll figure monetization later.” → “Here is our path to positive unit economics.” - “We’re building an AI co-pilot for X.” → “We automate Y% of this workflow with Z% error bound.” - “TAM is $100B+.” → “Here’s the niche we can own in 24 months.” - “Our model is better than OpenAI.” → “We’re closer to the data, workflow and distribution.” - “Trust us, it works.” → “Here is the eval suite and uptime history.” Share a tip: drop one metric or governance artifact you now expect to see in any serious AI deck. 👇 #AI #Startups #India #VentureCapital #Founders #ProductLeadership #DataDriven

Over the last few days, we've seen another wave of AI announcements. New models. New features. New AI assistants. New product launches. If you only looked at the headlines, you'd think the AI race is all about building the smartest model. But after spending years working with SaaS founders at DealMirror, I think the real shift is happening somewhere else. We're moving from the era of "AI-powered products" to the era of "AI-powered businesses." Having AI is no longer a differentiator. Every week, we see more products launching with AI features. The challenge isn't adding AI—it's making AI solve a real business problem. The questions buyers are asking have changed: Will this save my team time? Can I trust it with my data? Does it fit into my existing workflow? Will I still be using it six months from now? The winners won't be the companies with the longest feature list. They'll be the ones that create the most value after the first login. That's the trend I'm watching this week. If you're building a SaaS product today, are you competing on AI features—or on customer outcomes? I'd love to hear your perspective. #MarketplaceMonday #SaaS #ArtificialIntelligence #Founder #Startup #DealMirror

Rime's funding caught my attention for one tiny reason: pronunciation. TechCrunch reported that Rime raised a $24M Series A led by M13. The startup builds voice AI for enterprise calls, and one thing it focuses on is getting brand names and industry terms pronounced correctly. Source: https://lnkd.in/g5DDnw5X That sounds small until you imagine being the customer on the phone. If an AI voice gets the company name wrong, pauses in the wrong place, or says a normal term weirdly, you do not think, "wow, powerful model." You think, "this company put me inside a machine." That is the product detail I keep underlining. Adoption depends on both things: what the model can do, and whether the experience feels careful enough for a real customer. The demo can look impressive. The call can still feel careless. Especially in support, finance, travel, or any workflow with trust attached, small language details become product details. That is what I liked about this update. Rime is going after one of the tiny moments where trust leaks out. #Startups #VoiceAI #CustomerExperience

GrowthX® presents Sarvam Epoch Buildathon 200 builders. 8 hours. ₹3L in cash. Powered by Lightspeed India & Supported by Razorpay. The biggest opportunity isn't the prize money. The top 10 teams will present what they build at Epoch, India's Frontier AI Conference, where founders, investors, AI leaders, engineers, and decision-makers from NVIDIA, AWS, HCLTech, Zepto, Swiggy, Razorpay, and India's leading AI companies will be in the room. Most buildathons end with prizes. This one gives you the opportunity to demo at India's biggest AI product launch. About the Buildathon You'll spend 8 hours building voice, language, document intelligence, and multilingual AI products using Sarvam's AI stack. This Buildathon is your access to the Epoch stage. Who is this for? Founders building AI companies Engineers building models, agents & products Product leaders taking AI to millions of users Researchers & technical generalists Growth engineers & FDEs If you like shipping products and want to build alongside 200 serious builders, this room is for you Applications close in 48 hours. Request an invite using the link in the comments.

Everyone wants to build AI. But great products aren't powered by AI alone. They're powered by engineering that makes AI work—every single second. Here's how Zomato engineered ML at scale! 👇 #startups #AI #machinelearning #artificialintelligence #MLEngineering #MLOps #SoftwareArchitecture #SystemDesign #TechDeepDive #Engineering