Soorya Eswaran

Running LLMs on mobile phones—completely offline and locally—has become a reality. I recently tried the Google AI Edge Gallery app and was surprised by how effortlessly it brings generative AI to your device, no internet connection required after the initial setup. The app lets you chat, experiment with prompts, and even analyze images, all on your phone. If you're curious to try on-device GenAI for yourself, here’s the link: https://lnkd.in/dFJ9MXWV #GenAI #EdgeAI #LLM #MobileAI

35

2 Comments

18 Years of Vision: A Reflection on an Era at Adobe After 18 years at the helm, Shantanu Narayen has decided to step down as CEO of Adobe. 18 years is more than a tenure. It’s an era. When he became CEO in 2007, Adobe looked very different. Software came in boxes. Products were sold once. Updates arrived every few years. Then came a bold decision that many doubted. Adobe moved from boxed software to the cloud. Photoshop. Illustrator. Premiere Pro. All transitioned to Creative Cloud, turning a traditional software company into a subscription-driven platform. At the time, it looked risky. Today, it’s taught in business schools. During those years, Adobe didn’t just evolve its products. It reshaped its identity. Revenue grew from roughly $3B in 2007 to over $23B+ annually. Adobe expanded from creative tools into digital marketing and customer experience platforms. Its market value climbed from about $20B to hundreds of billions at its peak. But perhaps the bigger transformation was cultural. Adobe stopped thinking like a product company. It started thinking like a platform for creators and businesses worldwide. And now, as Narayen steps aside, Adobe stands at another turning point. Artificial Intelligence. Generative tools. Creative automation. New ways of building, designing, and telling stories. The next chapter will be written in AI. But the foundation of that chapter was built over the last 18 years. Leadership transitions are often described as endings. But the best ones feel more like passing the baton in the middle of a race. Because companies like Adobe don’t move in decades. They move in eras. And this was one of them. ​Shantanu Narayen's 18-year journey at Adobe is a testament to his exceptional leadership, vision, and dedication. He has transformed Adobe into a global powerhouse, and his impact on the creative and technology industries will be felt for years to come. As we pay tribute to his remarkable tenure, we can look forward to seeing what the future holds for Adobe under his continued guidance. The question now isn’t just who becomes the next CEO. It’s: What will the next 18 years of creativity look like? #Leadership #Adobe #Innovation #TechnologyLeadership #FutureOfWork

48

Shantanu Narayen’s 18 year tenure as CEO at Adobe is a reminder of how rare long-term stewardship has become in Silicon Valley. During his time with the company, the company’s stock increased more than 6x and its business model fundamentally changed. The most important decision of his tenure was moving Adobe from selling boxed software like Creative Suite to a subscription platform through Creative Cloud. At the time it was controversial. Today, it is considered one of the most successful SaaS transitions in the history of enterprise software, turning Photoshop, Illustrator, and Acrobat into one of the most profitable recurring software ecosystems ever built. The respect he commands across the industry says a lot about how difficult that transformation actually was. Microsoft CEO Satya Nadella even posted publicly congratulating Narayen on a “legendary run,” praising both his leadership and the way he expanded what creators and businesses could do with software. In an industry that is often intensely competitive, that kind of recognition from peers is not given lightly. There are only a few CEOs in the world who truly get to define one era of software and then help set the stage for the next. Narayen did it by turning Adobe into a SaaS powerhouse. However, the next leader will have to answer a harder question. What does a creative software company look like when AI becomes part of the creative process itself? #Adobe #Leadership #SaaS #AI #Technology

10

DC-DC converter efficiency principles just saved me ₹50,000 in ad spend. Here's how engineering logic crushed marketing waste. Last month my Meta campaigns were bleeding money. ₹2.5 lakh budget. ₹80 cost per lead. Sound familiar? ↳ 𝗧𝗵𝗲 𝗮𝗱 𝘀𝗽𝗲𝗻𝗱 𝘄𝗮𝘀𝘁𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺 Your campaigns run like inefficient power supplies. High input. Low output. Maximum heat generation. Most marketers accept 60% efficiency rates. They think conversion loss is normal. Budget goes in at 100% Platform fees eat 15% Wrong audience targeting wastes 25% Poor creative performance loses 20% Bad landing pages kill 15% You end up with 25% actual results. The rest becomes expensive heat. ↳ 𝗕𝘂𝗰𝗸 𝗰𝗼𝗻𝘃𝗲𝗿𝘁𝗲𝗿 𝘄𝗶𝘀𝗱𝗼𝗺 𝗮𝗽𝗽𝗹𝗶𝗲𝗱 Power engineers never accept such waste. Buck converters achieve 95% efficiency rates. They do three things right. ☑ Control switching frequency precisely ☑ Minimize resistance in current path ☑ Reduce energy loss at each transition Your ad campaigns need the same discipline. Every rupee must flow with purpose. Every click must convert with intention. ↳ 𝗖𝗮𝗺𝗽𝗮𝗶𝗴𝗻 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗽𝗮𝗿𝗮𝗹𝗹𝗲𝗹 I treated my Meta ads like a power circuit. Input power equals ad budget. Output power equals qualified leads. Efficiency equals leads per rupee spent. Started measuring conversion at every stage. Impression to click efficiency Click to landing page efficiency Landing page to lead efficiency Lead to customer efficiency Found the biggest energy drain. Wrong audience segments were burning ₹1,200 daily. ↳ 𝗧𝗵𝗲 ₹𝟱𝟬,𝟬𝟬𝟬 𝘀𝗮𝘃𝗶𝗻𝗴𝘀 𝗿𝗲𝘀𝘂𝗹𝘁 Applied power supply optimization logic. Reduced switching losses by tighter targeting. Minimized resistance through better creative flow. Eliminated heat generation via audience precision. 𝗥𝗲𝘀𝘂𝗹𝘁𝘀 𝗮𝗳𝘁𝗲𝗿 𝟯𝟬 𝗱𝗮𝘆𝘀. Cost per lead dropped from ₹80 to ₹35. Same budget. 2.3x more qualified leads. Monthly savings crossed ₹50,000. 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗽𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲𝘀 𝘄𝗼𝗿𝗸 𝗶𝗻 𝗺𝗮𝗿𝗸𝗲𝘁𝗶𝗻𝗴 𝘁𝗼𝗼. You just need to think like a circuit designer. What's the biggest energy drain in your current campaigns?

13

The news of a $100,000 H-1B visa renewal fee feels incredibly harsh for Indian professionals in the U.S. At first, it looks like another “shock” tactic — something extreme that eventually gets bargained down. We’ve seen this before: restrictions on AI chip sales to China that later turned into a 15% tax. But even as a proposal, this sends ripples across the industry: 🌍 Short term: Companies may accelerate offshoring. That could open more opportunities for freshers in India, and for students in U.S. universities as firms look for alternatives to costly renewals. 💻 Reality check: A large share of the U.S. tech workforce is Indian or Chinese. At places like Amazon and Uber, I noticed a pattern — the tougher the environment, the more you see immigrants. Not because it’s easier for them, but because survival and resilience are part of their journey. ⏳ Long term: America once proudly said “immigrants built this country.” Now the narrative feels closer to “protect local jobs.” If that mindset hardens, the U.S. risks losing its edge as top talent may simply choose Canada, Europe, or Asia instead. This also reminded me of something from my civics classes: India’s founding leaders debated whether to adopt a presidential system. The planning committee (led by Ambedkar and others) felt too much power in one individual could be dangerous. As a student, I believed a presidential system might have enabled faster decisions. Today, I see how that foresight may have guarded us from concentrated power — something we see tested in democracies worldwide. Whether it’s global tech policy or political systems, checks and balances matter more than speed. 👉 What do you think — will this visa fee push more roles offshore, or are we witnessing a bigger reset in how global talent and power structures are managed? My own perspective has shifted. I once believed the U.S. would always welcome immigrant talent. Today, I’m not so sure. What do you think — will this just push more roles offshore, or are we watching a bigger reset in how global tech talent is sourced? #H1B #Immigration #FutureOfWork #USIndia #TechTalent

10

1 Comment

🇮🇳🏆 821 million views. 72 million concurrent connections. One legendary trophy and One interesting #AgenticAI based Engineering problem. While celebrating India’s victory in #T20WorldCup2026 against New Zealand last night, my mind stumbled upon a critical tech problem. Imagine a scenario - #JioHotstar rolls out a new feature called vertical live feed and a 360-degree multi-cam option during this Final (hypothetical scenario). The stream starts buffering, millions of fans start switching camera angles, and event data is pouring directly into the analytics engine and all of a sudden all the viewers start getting a lag in stream. 🚨 For such crucial services and scenarios, we engineers always ensure the hygiene stuff like., horizontal scaling, circuit breakers, compartmentalisation, proper resource provisioning and other edge cases. What remains unsolved is the "unknown unknowns." For those, only logs remain. When something at this scale breaks on production, looking at grafana and cloudwatch for logs & dashboards becomes a nightmare. Dashboards help establish that there is an issue, but pinpointing the exact Root Cause Analysis is incredibly difficult. #AgenticAI might help in solving this by replacing static dashboards with an autonomous execution. Lets see how ⚡ The Trigger and State Management The moment a latency anomaly spikes and stays for a period of time, the backend triggers an investigation payload. Which kicks off a Temporal activity. here #Temporal can provide the durability needed during autonomous RCA pipeline. 🧠 The Memory Retrieval Temporal wakes up the LangGraph orchestrator (Alpha). It first queries #Mem0 (plus analytics store), the long-term memory layer to fetch similar bitrate incidents that happened in past and their exact causes. 🕵️‍♂️ The Cognitive Delegation Armed with context, Alpha routes the raw logs to the Analytics Agent (Zen). Powered by Claude Sonnet 4.6, Zen translates the vague anomaly into a highly optimised Python script designed to crunch millions of click-stream events and isolate the specific buffering devices. 🛡️ The Secure Execution Running AI-generated code on production servers is dangerous. LangGraph drops Zen's code into a #Docker container secured by #gVisor, which acts as a strict user-space kernel, executing the code in total quarantine. This ensures a hallucinated line of code does not bring down the core streaming infrastructure. Note: For heavier log processing workloads, Zen can submit a #Spark job as well for distributed processing. 🔄 The Resolution Loop LangGraph continuously observes this ReAct loop. Once Zen securely identifies the bug, Alpha consults Dash, the growth agent, to calculate the exact user drop-off impact. Only then does the system autonomously push the hotfix for the human review. Its just an initial thought, lets discuss further in comments 👍

21

Got a panicked WhatsApp at 7 AM this morning. "Our payment gateway is rejecting everyone. We've lost ₹8 lakhs in sales overnight." Not a technical glitch. Human error caused by "smart" automation. Here's what happened: This jewelry brand does ₹25 lakhs monthly. Last year, they hired a fintech consultant who said: "You're losing money to fraud. Let me build you an intelligent fraud detection system." Cost: ₹6.5 lakhs. The system included: → AI-powered risk scoring → Real-time transaction analysis → Behavioral pattern detection → Automatic blocking of suspicious orders → Machine learning from past fraud attempts → Integration with 3 payment gateways Sounded necessary. He showed them the fraud statistics. Scared them properly. They signed. First month: Blocked 12 transactions. All were actual fraud. Great! Second month: Blocked 45 transactions. 8 were legitimate customers. Not great. Third month: Blocked 180 transactions. 140 were real customers trying to buy. The AI decided: → Orders above ₹50,000 = Suspicious (even from repeat customers) → Using different delivery address = Fraud risk → Ordering late night = Red flag → New customer + high value = Auto-reject Real customers got declined. Called customer service, furious. "Your site rejected my card!" "It's a gift, why can't I ship elsewhere?" "I work night shifts, am I not allowed to shop?" The consultant? "AI is being cautious. That's good. We can fine-tune the sensitivity for ₹1 lakhs more." They called us instead. We rebuilt it in 6-7 days with simple rules: → Flag orders above ₹1 lakh for manual review (not auto-reject) → Allow different shipping for orders below ₹50K → Verify via OTP, don't just block → Let customers talk to humans instantly if flagged → Track patterns but don't auto-punish Results in 90 days: → Actual fraud caught: Still 100% (12 attempts blocked) → False positives: 140 per month → 3 per month → Revenue recovered: ₹8L+ monthly → Customer satisfaction: Back to normal → Cart abandonment: Down 52% The owner told me yesterday: "I was so scared of losing ₹50,000 to fraud that I lost ₹6+ lakhs to paranoia." Here's the lesson: Automation that blocks customers is worse than no automation. Fraud protection should protect your business. Not kill it. AI shouldn't make final decisions on revenue. Humans should. Flag. Alert. Verify. Don't just reject and hope the AI is right. Because every blocked real customer is a lost customer forever. Ever been blocked by a system that thought you were a fraud? #AIAutomation #EcommerceBusiness #FraudPrevention #BusinessAutomation #StartupLessons #PaymentSecurity #Entrepreneurship #TechSimplicity #SmallBusiness #AITools #StartupGrowth #WorkflowAutomation #n8n #DubaiBusiness #EcommerceTech

11

2 Comments

When I lived in Hyderabad, one of my favorite restaurants was Chutneys. When they opened a branch in Inorbit Mall, I was genuinely happy because it was closer to where I lived. What surprised me was the scale of their waiting area. In an expensive mall like Inorbit, Chutney’s had invested in a large, comfortable waiting space with luxurious sofas, almost indistinguishable from the actual seating area. I remember thinking the investment must have been huge, but it made the experience feel effortless. People could wait without feeling frustrated, even during peak hours. I experienced something similar in a very different context a year or two ago, while admitting my child to school. For some reason, we reached the admission venue two hours early on a weekday. Only after arriving did we realize we had forgotten our documents at home. We rushed back, we got the folder, and everything worked out. Had we arrived just in time, we would have missed the admission entirely. Given how selective the school was, that would have meant losing a carefully chosen opportunity or creating a lot of avoidable stress. Much later, while reading Thinking in Systems, these two incidents suddenly connected. Donella Meadows talks about buffers as one of the most powerful but often misunderstood elements of a system. A buffer is simply the size of a stabilizing stock relative to its flows. Time, inventory, space, patience, slack, and margin are all buffers. Buffers come at a cost. Maintaining extra capacity, time, or inventory always feels inefficient. Reaching two hours early meant taking time away from something else. Chutney’s waiting area meant higher real estate and furnishing costs. But buffers exist precisely to absorb variability and uncertainty. They prevent systems from failing when reality does not behave as planned. Buffers also reduce oscillations. For example, manufacturing systems that respond to rising demand by aggressively increasing orders, only to be overwhelmed with excess inventory when demand falls. Without adequate buffers, delays and overreactions amplify swings. With buffers in place, the same system becomes calmer, more predictable, and far less stressful to operate. Whether in supply chains, restaurants, or our personal lives, buffers act as shock absorbers. They smooth fluctuations, reduce anxiety, and prevent small mistakes from cascading into major failures. Not everyone can afford large buffers all the time, but Thinking in Systems makes one thing clear: systems without buffers are brittle. Sometimes, all you need is a little margin.

56

6 Comments

💡 How A $20 AI Subscription Saved $160,000 A man in the US recently shared how he used an AI chatbot to bring down a $195,000 (INR 1.6 crore) hospital bill to $33,000 (INR 27 lakh) all thanks to Claude, an AI tool by Anthropic. When the hospital billed his late brother-in-law for just four hours of care, he used AI to analyze every line item, uncover duplicate charges, wrong billing codes, and even violations of medical norms. With AI’s help, he drafted firm, well-reasoned letters and the hospital eventually agreed to cut the bill by over 80%. His takeaway: “My $20-a-month AI subscription was the best investment I ever made.” Stories like this remind me that AI isn’t just for techies. It’s for everyone who wants to think smarter, save time, and take back control. That’s why I keep telling people, start incorporating AI tools into your day-to-day life. #AI #Claude #Anthropic #ArtificialIntelligence #Productivity #ConsumerEmpowerment #Innovation #AIForthePeople

41

3 Comments

The $100,000 question isn't just about visa fees – it's about where the future of AI talent will thrive 🚀 The H1B drama on Twitter? It's getting REAL. 📍 Visa fees jumped from $5,000 to $100,000 overnight 📍 Microsoft told Indian employees: "Don't leave the US. Stay put." 📍 Trolls launched "Operation Clog the Toilet" to block Indians' flight bookings Harsh? Yes. But there's a bigger story here. **Will Indians actually return home?** 🤔 The reality check: • Panic bookings before the deadline? ✓ • Most H1B holders staying put? Also ✓ • The real exodus? 50,000+ laid-off techies with no choice But wait – India's quietly building something interesting. Tamil Nadu is offering competitive packages to lure scientists back. 60% of global AI startups have Indian founders. We've got the PARAM Siddhi supercomputer. IITs are building AI research clusters. Even Zoho's founder is calling techies home: "Do not live in fear. Make the bold move." 💪 Here's the plot twist: The AI era doesn't care about your visa status. 94% of Indian professionals are already using AI tools. Companies don't need more people – they need AI-skilled people. Master AI, and you'll thrive in Silicon Valley OR Bengaluru. We're not seeing brain drain anymore. We're seeing brain circulation 🔄 Remote work + AI tools + global teams = Geography matters less than capability The uncomfortable truth? When companies pay $100,000 for a visa, they're not buying a nationality. They're buying irreplaceable talent. So here's my question: Will this visa chaos accidentally turn India into an AI powerhouse? Or will talent always find its way to Silicon Valley, no matter the cost? Drop your thoughts below 👇 #H1BVisa #AITalent #TechIndustry #BrainDrain #IndianTech #FutureOfWork

5