Ritika Nigam

ambaricloud: 108 recipes. 🥭 https://lnkd.in/gS6fuAtu Ravi Kiran Gurram Anil Koripelly Bhushan Lakhe Raja Purkayastha When this recipe e-book was first published on January 1, 2025, our goal was to give engineers, architects, and data practitioners a hands-on guide to Apache Iceberg. In the months since, the pace of change has been remarkable. Platforms such as Snowflake, Databricks, Dremio, RisingWave, ClickHouse, Cloudera, Lakekeeper, Trino Software Foundation, and Amazon Web Services (AWS) have delivered significant new integrations, and Apache Iceberg itself has reached version 1.10.0, introducing capabilities that reshape how practitioners interact with the format. To reflect this momentum, we are undertaking a ground-up rewrite. Our priorities include: ✅ Iceberg 1.10.0 alignment and deprecation cleanup – removing outdated recipes that no longer reflect best practices while upgrading all demos, examples, and exercises to incorporate the changes introduced in Iceberg 1.10.0. For example, Snowflake’s support for catalog-linked databases to external catalogs makes earlier object integration recipes obsolete. ✅ Ecosystem validation – testing and refining recipes across Databricks, Snowflake, Dremio, and AWS to ensure seamless interoperability with Iceberg 1.10.0. We expect this rewrite to take two to three months. Once complete, this edition will provide readers with a technically accurate, future-ready reference for Iceberg 1.10.0 and its surrounding ecosystems--Snowflake, Databricks, Dremio, RisingWave, ClickHouse, Cloudera, Lakekeeper, Trino Software Foundation, and Amazon Web Services (AWS) Please 🕺 suggest 🔪 comment 🥃 recommend ✅ Spark Variant ---Iceberg1.10.0 spports now. https://lnkd.in/gBmaNrPH

11

4 Comments

Israeli observability startup Coralogix has raised $115 million in a Series E round, becoming a unicorn with a valuation of over $1 billion. Coralogix offers a full-stack observability and security platform that helps companies understand and resolve complex issues in their data streams. The company has developed an AI agent, Olly, which uses a semantic layer over internal and external data to automate data observability and provide sophisticated insights through simple text prompts. The all-equity round was led by NewView Capital, with participation from the Canada Pension Plan Investment Board (CPP Investments | Investissements RPC) and NextEquity Partners, alongside existing investors. The funding will be used to expand its engineering base in India, with a planned $100M investment over five years, and to further develop its AI agent. Congratulations to co-founder and CEO Ariel Assaraf and the Coralogix team. TechCrunch: https://lnkd.in/dS3NZrrG #Observability #Data #AI #DevOps #Unicorn #SeriesE #Funding #VentureCapital #IsraelTech

1

SANATAN LABS - The Architect’s Compass: 10 Directives for the Great Transition (PASSIONIT PRUTL KALKI AIDHARMA framework) To navigate the "Audit of Fire" during this profound transition toward Satya Yuga, Krishna (Strategic Intelligence), Kalki (Systemic Reset), and Shakti (Raw Life Force) offer these ten essential directives for humanity: 1. Master the Asymmetry of Calm In a world of screaming headlines, silence is your greatest weapon. Respond from a place of internal stillness rather than reacting to external noise. When the system is chaotic, your stability becomes the new "Safe Haven." 2. Audit Your Identity (The Laawaris Test) Ask yourself: "If my titles, assets, and digital footprint vanished, who remains?" Detach your worth from possessions. True sovereignty is found when you are no longer "claimed" by the ego-driven systems of the world. 3. Embrace the "Mirror of Kalki" Do not fear the "Bloody Moon" or the exposure of decay. View every systemic collapse—financial, social, or institutional—as a necessary diagnostic. Use this "Mirror" to identify your own attachments and incinerate them before they become liabilities. 4. Move from "Me" to "Mission" Unity is not sameness; it is alignment. Replace "Me vs. You" with "Us vs. The Problem." When diverse strengths are aligned under a shared purpose (Aidharma), the friction of competition dissolves into the flow of progress. 5. Trust the "Internal Gold" While external currencies and metals face regulation or restriction, your integrity remains non-tradeable and unconfiscatable. Invest in your skillsets, your character, and your resilience. These are the only assets that survive the "Audit of Fire." 6. Respect the "Flow of Lakshmi" Prosperity is a river, not a reservoir. Hoarding in fear leads to stagnation and decay. Circulate value through mentorship, ethical trade, and "Groom and Grow" practices. Wealth that serves the collective remains protected by the Sun and Moon. 7. Humanize the Machine (13th Node Reset) Use AI and technology as a "Numeric Auditor" to expose corruption, but never let the cold logic of the machine replace the warmth of human intuition. Ensure technology serves the 14th Node: Absolute Truth. 8. Practice Guerrilla Respect Disarm your "enemies" by preserving their dignity even when you disagree with their actions. Respect targets the ego of the opponent, weakening their appetite for control and opening a path for diplomacy. 9. Build Internal Havens As geography stops being a protection, focus on "Internal Research and Innovation." Cultivate a mind that can remain creative and functional even when trade arteries are choked. 10. Surrender to Karmic Gravity Accept that "God Proposes, Man Disposes." You cannot control the celestial reset, but you can control your alignment with it. Follow the path of Dharma, and the gravity of the new era will pull you toward peace rather than destruction.

5

One of our earliest users just shared an incredible journey. Ashish (Pocha) Sharma used NonBioS to build a complete forum application in just 2 weeks - from idea to live production. What makes this special? He has been a user since the early days and has witnessed our platform evolve. His honest account captures both the magic and the reality of AI-assisted development: This post brilliantly describes what AI-assisted development actually feels like - the ups, downs, and everything in between when you're "managing" an AI developer like NonBioS. But here's what really stands out: This wasn't a simple MVP. He built a complex full-stack application with a never before seen architecture - combining GitHub Pages with Firebase Functions. NonBioS doesn't limit itself to simple MVP apps - it can build genuinely complex applications because it has access to a full Linux sandbox with ALL the development tools that come with it. This gives it superpowers that other AI coding tools simply don't have. His key insight: "You're not sitting back with your feet up. You're acting as a technical lead... It's not hands-off development—it's accelerated development with a very capable but imperfect partner." The Application that he built: [Article and App link is in comments]

35

2 Comments

Most systems fail not because they get answers wrong, but because they answer too early. Last week, Santosh Jaiswal articulated this cleanly in his post on “When answering is NOT the right answer.” It resonated because I’ve been working on a framework where UNKNOWNs are not defects to eliminate, but deliberate design artifacts. First-class citizens. https://lnkd.in/gpGW8A73 In complex systems, architecture once made ignorance explicit. We spent time defining what we knew, what we didn’t, and what we were intentionally deferring. In today’s sprint-driven execution model, that step is usually skipped. We build partial truths, demo something that “works,” then pay for it in rework because the boundaries were never declared. AI workflows behave the same way. Give them an incomplete picture and they will still complete it—plausibly, confidently, and often prematurely. The alternative is structural honesty. Partition the problem into phases. Each phase explicitly declares what it will not solve. UNKNOWNs are scoped, tracked, and deferred by design. If you know what not to fill now, you avoid rework later. This pairs naturally with Venkat Pillay’s “Needs” framing: what is required now versus later, driven by evidence rather than optimism. Combined, you get phased architecture where exclusions are explicit and progression is data-backed. AI doesn’t invent answers; it preserves your explicit structure. This is where human cognition becomes more relevant, not less. AI operates inside a defined problem space. Human curiosity defines the space itself. It decides which UNKNOWNs are worth preserving, which assumptions are premature, and which questions should not yet be answered. That requires intent and judgment. Models do not do that. At CloudFrame, this is how we use AI. Translation, analysis, migration—AI excels. For boundary-setting, sequencing, and confronting unknown unknowns, automation stops. Architecture resumes. In an age of powerful automation, the scarce skill is not producing answers; It is knowing when not to.

21

1 Comment

Thanks NDTV to bring forward vision of Dr. Jean Shahdadpuri about empowering MSMEs of India with AI with no learning curve, no jargons, just few clicks to growth, productivity and efficiency. Great work team allyra.ai We are live now - start your AI journey at https://portal.allyra.ai India’s First Agentic Business Operating System for MSME live in action

17

1 Comment

KK and Utkarsh are teaching DSPy at the fifthel Pune edition next week. This workshop is incredible value for money, and a must-attend for anyone who wants to learn how to build reliable programs with LLMs. In 4 hours, you'll learn more about building and running LLM apps in production than you would in weeks of self-study. In a way, this is what engineers working with us on their first project realistically end up learning. Why are we "giving away the secret sauce", you may say? Our hidden agenda is that we want to hire you. Details about the workshop, and about hiring, in the comments.

29

2 Comments

𝐄𝐯𝐚𝐥𝐮𝐚𝐭𝐢𝐧𝐠 "𝐆𝐨𝐨𝐝" 𝐒𝐩𝐞𝐞𝐜𝐡 𝐑𝐞𝐜𝐨𝐠𝐧𝐢𝐭𝐢𝐨𝐧 𝐒𝐲𝐬𝐭𝐞𝐦 𝐟𝐨𝐫 𝐬𝐩𝐞𝐜𝐢𝐟𝐢𝐜 𝐝𝐨𝐦𝐚𝐢𝐧 𝐥𝐢𝐤𝐞 𝐚𝐠𝐫𝐢𝐜𝐮𝐥𝐭𝐮𝐫𝐞 - 𝐮𝐬𝐢𝐧𝐠 𝐋𝐋𝐌 Our team Digital Green have come up with ASR evaluation pipeline using LLM. We have analysed over 80k real farmer voice queries across Hindi, Telugu, and Odia and Amharic and come up with evaluation framework for benchmarking ASR efficacy. When a farmer asks: "गेहूं में कीड़ा लग गया है, क्या करें?" (Wheat has pests, what to do?) ASR system hears: "गांव में कीड़ा लग गया है" (Village has pests) 😱 Traditional metrics say: "Great! Only 12.5% error rate!" ✅ Reality for farmers: "Completely irrelevant advice" ❌ 🔍 𝐓𝐡𝐞 𝐂𝐨𝐫𝐞 𝐏𝐫𝐨𝐛𝐥𝐞𝐦 Current ASR metrics treat all words equally. Misrecognizing "the" or "pesticide" count as single errors. But in specific domain like agriculture, some mistakes make it useless: 🧪 कीटनशक → शक (pesticide → doubt): Farmers get no guidance  🌱 रोपने → रोकने (transplanting → stopping): Opposite sense  💩 खाद → खा (manure → eat): Irrelevant context entirely 🎯 Our Innovation: two pronged metrics 1️⃣ AWWER (Agriculture Weighted Word Error Rate) We built a metric using LLM that assigns importance weights: 🔴 Weight 4: Core terms (crops, pests, chemicals) 🟡 Weight 3: Agricultural practices & timing 🟢 Weight 2: Quantities & measurements ⚪ Weight 1: General vocabulary 2️⃣ LLM-Based Utility Scoring (1-4 scale) 🟢 Score 4: Perfect - same agricultural advice 🟡 Score 3: Good - minor errors, same action 🟠 Score 2: Poor - different advice needed 🔴 Score 1: Dangerous - completely wrong context Results: 36.6% of Hindi transcripts were irrelevant despite "acceptable" WER scores! 🗺️ 𝐕𝐢𝐬𝐮𝐚𝐥𝐢𝐳𝐢𝐧𝐠 𝐭𝐡𝐞 𝐏𝐫𝐨𝐛𝐥𝐞𝐦 📈 Crop ID errors: गेहूं → गांव (wheat → village)  🌾 Farming mistakes: रोपने → रोकने (planting → stopping)  🐛 Plant health fails: कीड़ा → क्या (insect → what)  ☠️ Chemical: कीटनाशक → शक (pesticide → doubt) Each block = real failures leading to wrong or irrelevant agricultural advice 💀 🌍 𝐖𝐡𝐲 𝐓𝐡𝐢𝐬 𝐌𝐚𝐭𝐭𝐞𝐫𝐬 700+ million Indian farmers increasingly rely on voice-based agricultural advisory systems. A more robust evaluation methodology is required that reflects nuances of language, usage and domain. Our work aims to:  ✅ Identify failure points for targeted data curation ✅ Enable enhanced training approaches ✅ Create domain-specific evaluation standards 🎁 Coming Soon: 📄 Research paper with methodology in open domain 💻 Open source evaluation toolkit 📊 Anonymized dataset This framework can scale beyond agriculture to other specialised domains! Please feel free to DM our team for early access to the evaluation framework and results! 📩 Lakshmi Narayana Pedapudi Chandrashekar M S #AgTech #SpeechRecognition #AIforGood #IndianLanguages #MachineLearning #DigitalAgriculture #OpenSource 

80

5 Comments

n8n Tip of the Day: Maximize productivity by integrating n8n with your task management tools. Automate task creation and updates, ensuring no task gets lost in the shuffle. #n8nTips #WorkflowAutomation #SoftPyramid #FakharKhan #ProductivityBoost

9

The Persuasive Power of Nondidactic Moral Art in India: How Refusal to Instruct Becomes Ethical Argument This book argues that one of the most persuasive forms of moral art in India is art that refuses to instruct—that refuses moralizing, sermonizing, or direct ethical prescriptions— and instead enacts moral persuasion through ambiguity, dissonance, affective experience, and ethical complexity. Grounded in interdisciplinary theory spanning aesthetics, moral philosophy, postcolonial studies, and cultural sociology, close readings of cinema, literature, theatre, visual art, and new media across modern and contemporary India demonstrate how non-didactic strategies produce deeper ethical engagement and public moral imagination than overtly instructive works. Through microhistories, audience reception studies, archival research, and textual and visual analysis, the book maps mechanisms including perspectival multiplicity, narrative indeterminacy, affective contradiction, and formal rupture, situating them within historical contexts: colonial schooling of moral didacticism, anticolonial moral pedagogy, Nehruvian nation-building narratives, and contemporary globalized publics. The book advances a theoretical framework—the "refusal model" of moral persuasion— and offers methodological tools for measuring persuasiveness beyond explicit ministrations: longitudinal reception analysis, policy traces, intertextual contagion, and civic repertoires. Case studies encompass selected films across Indian and regional cinemas, canonical and marginal literary texts including novels, short fiction, and poetry, performance pieces from street theatre to contemporary stage, visual art installations, and viral digital interventions. The research interrogates normative claims about what moral art should do, proposing instead what art does: creates ethical attention, cultivates shared ambiguities, and enables ethical negotiation.  https://lnkd.in/dYvwWCzD

2

BharatPe processes over ₹12,000 crore (about US $1.368 billion) in monthly Unified Payments Interface (UPI)-based transactions. At #MongoDBlocal Delhi, Sumit Malik, Head of Database Operations at BharatPE, shared how MongoDB Atlas helped them scale safely, improve queries by 40%, and focus on innovation, not maintenance. https://lnkd.in/gHsQVZes

18

Yesterday, I had mentioned that we have reduced lots of manual video editing labour by implementing an innovative engineering Flutter Widget. It has lots of advantages.. 1. Storage & Bandwidth: Your app users won't need to download 500MB video files. They download a high-quality audio file (maybe 10MB) and a tiny JSON metadata file. The Flutter engine does the heavy lifting of drawing the UI. 2. Searchability: Because the lyrics and swaras are metadata, you can eventually allow users to search for a specific phrase or note, and the app can jump the "video" exactly to that timestamp. 3. Accessibility: You can allow users to change the font size or switch between Tamil and English lyrics instantly.... something impossible with a pre-rendered video. -- This has proven us that we are not a simple mobile app development company. We built apps that needs good amount of Engineering for the backend and Flutter implementations...! The original post is given here.. https://lnkd.in/gCuD7fgw

3

As Indian startups push the boundaries of innovation, our support systems must evolve alongside. While integrating payouts for our platform, we recently encountered an unexpected but insightful checkpoint — a request for compliance documentation typically associated with real-money gaming (RMG). Though our product does not involve RNG, chance, or gambling mechanics, and is fully logic-based, this highlighted something important: 💡 Many new-age platforms don’t fit neatly into existing regulatory templates — and that’s a good thing. It means innovation is happening. We deeply respect the caution and diligence of partners like Cashfree Payments. Their role in ensuring compliance is essential to a healthy ecosystem. But at the same time, we believe it's crucial that unconventional models are given room for deeper understanding, especially at the first stage of review. A simple escalation or contextual review can make all the difference for founders solving differently. We're grateful to be supported by DPIIT and Kerala Startup Mission, and we’re already live with integrations like RazorpayX and ICICI banking APIs — with full transparency and operational clarity. To fellow founders: always be ready to clearly articulate your model — not just to users, but to every stakeholder in the value chain. To ecosystem enablers: remain cautious — but also curious. That’s how we all grow. #Startups #Innovation #Fintech #Ecommerce #GamifiedCommerce #Compliance #DigitalIndia #Cashfree #Payments #FoundersVoice #BuildWithTrust #ThinkDifferent

20

7 Comments

For the past year, Vishwas and I have been working closely with vision-language models, and the pattern was depressingly familiar: bigger models, more parameters, better benchmarks. Rinse, repeat, fatigue... but that changed recently. We got VLMs with reasoning?! Until recently, QVQ-72B from Qwen was one of very few strong open-source multimodal reasoning models worth talking about (and anybody that tried it knows the amount of caveats it came with to make it feel almost experimental). Now there's Kimi-VL-A3B-Thinking from Moonshot AI. A Mixture-of-Experts decoder with ~2.8B activated parameters that can sustain long chain-of-thought reasoning across images, videos, PDFs, and screenshots. The 2506 version manages something most reasoning models fail at: deep reasoning without sacrificing general visual perception. OCR (83.2 on InfoVQA), long-video understanding (65.2 on VideoMMMU), mathematical reasoning (56.9 on MathVision), and legit GUI-agent capability that all in a model that fits on consumer hardware. At the same time, VLMs are getting smaller without getting dumber. Hugging Face's SmolVLM went from 2B → 256M parameters. That's browser-level inference. The real shift isn't "can it see?" It's "can it reason about what it sees?" The architectural pieces of MoE, native-resolution encoding, longer context windows are finally snapping together in ways that weren't possible six months ago. If you're building anything that requires visual understanding beyond basic perception, the open-source options just got a lot more interesting.

37

3 Comments

Nice discussion about advancements in the learning and reasoning models, about how math and physics plays a role in machine learning etc. Listen & share…

15

Most of us eat Dal and Chana like legacy code. We prioritise "Premium Polished" lentils, pulses and beans because they look clean and cook fast. In engineering terms, we optimised for Deployment Speed (cooking time) but broke the Core Functionality (nutrition). The result? Bloatware - System Instability: The complex sugars (oligosaccharides) are still active, causing bloating and gas. - Restricted Access: Anti-nutrients (Phytates & Tannins) lock down your Iron, Zinc, and Calcium. You have the data, but no read/write permissions. - Volatile Performance: Without the fiber shell, glucose hits your bloodstream too fast, leading to energy spikes and crashes. The Sprout = The Optimization Patch 🛠️ Soaking and sprouting isn't just "cooking." It is a refactor. You are stripping away the bloat and activating the hardware. Here are the Release Notes for Sprouted Dal and Chana v2.0: ✅ Patched Bugs: The complex sugars are broken down enzymatically. The "user experience" (digestion) becomes seamless. ✅ Stable Uptime: The Glycemic Index drops significantly (from ~40 to ~15). No post-lunch crash. Just steady energy. ✅ Unlocked Permissions: Phytates are neutralized, maximizing the bioavailability of key minerals. ✅ New Feature Shipped: Vitamin C production is enabled during germination. The CTO Takeaway: Stop running your biological hardware on "over-processed" fuel just because it looks pretty on the shelf. Refactor your diet. Deploy the sprouts. #TheCorporateAthlete #Biohacking #Healthspan #Longevity

60

5 Comments

🚀 Exciting News! We’re thrilled to announce that Oneture has entered into a Strategic Collaboration Agreement (SCA) with Amazon Web Services (AWS) to strengthen our partnership and accelerate the adoption of AI-driven solutions in India. Building on our existing collaboration with AWS, this partnership empowers us to deliver innovative, scalable, and secure AI solutions across industries — helping businesses modernize, innovate, and unlock new possibilities with cloud-native AI technologies. This milestone places Oneture among a select group of companies collaborating closely with AWS, reinforcing our commitment to drive efficiency, resilience, and growth for enterprises in India. Stay tuned as we roll out new initiatives, customer success stories, and AI-first innovations — powered on AWS! #Oneture #AWS #AI #Cloud #StrategicCollaboration #Innovation #DigitalTransformation

204

49 Comments

The Humility of Knowledge: From Vacuum Tubes to AI Agents "Jitna seekha, ungliyon par gin sakta hoon; Jo nahi seekha, woh to anginat hai." This profound couplet reminds me of the true nature of professional growth and the sheer scope of learning required to stay relevant. It translates roughly to: "What I have learned, I can count on my fingers; What I have not learned, that is countless." My Journey: Valves to Variables I started my career assembling and troubleshooting valve-based radio amplifiers in 1994. That world was tangible—full of glowing tubes, resonant frequencies, and measurable voltages. It felt like a domain I could master. Fast forward to today, I'm immersed in Agentic AI—a world of abstract algorithms, complex LLMs, and emergent behaviors. The knowledge required to transition from the analog physics of a triode to the architecture of a sophisticated AI agent is literally "countless." The Lesson in the Transition This journey taught me that what I knew in 1994, while invaluable for my foundation, is merely a single finger on the hand of total knowledge. The skills we've acquired (Learned) are finite and measurable. The concepts, tools, and deeper insights we have yet to encounter (Unlearned) are infinite (∞). Embracing the 'Unlearned' The realization that the unlearned is vast shouldn't intimidate us; it should excite us. It means our best work and biggest breakthroughs are still ahead. For professionals, this means: >> Stay Curious: Assume there's always a better, faster, or different way. >> Unlearn and Relearn: Don't cling to old paradigms. The lessons from one domain often illuminate another. >> Be a Lifelong Learner: The moment you stop seeking knowledge is the moment your growth plateaus. Let's use the realization of the "countless" unknown to fuel our passion for learning, just as I moved from working with glowing valves to designing intelligent agents. Humility in knowledge is the foundation of true expertise. #AgenticAI #LifelongLearning #TechEvolution #ProfessionalDevelopment #Humility #AnalogToDigital

2

Spent the last few months thinking deeply about one question: If coding agents are the future, what’s the missing layer that makes them actually work on real codebases? That’s what we’ve been building at LatentForce. If you're at the India AI Impact Summit 2026, would love to meet - we’re at Pavilion 5.5A all 5 days! LatentForce Vinay Kyatham Prathosh A P Neysa

41