25Years

Today, Hema and I celebrate 25 years together—a true milestone that demands a moment of pause, deep reflection, and joyful celebration. Looks like 10th anniversary was just the other day! https://photos.app.goo.gl/w4LsSzZgbuc4u5mU6  25 Years: A Beautiful Pause for Reflection. 

Over this quarter-century, we’ve moved across countries, faced countless challenges, and created a lifetime of shared memories. The constant that always stands out is this: through every "thick and thin," we have remained fundamentally united, standing side-by-side.

Instead, marriage is an active creation. It’s the continuous process of re-creation, nurturing, growing, lifting, loving, giving, and supporting each other—every single day. I genuinely believe that being a great leader in the world starts right within the four walls of our home.

You, Hema, is the core backend resilient service to enable the stability and success of my frontend visible career, publications and academic life.  I couldn't achieve with alternative - it's only by your sacrifice. 

Our relationship is a dynamic co-creation, and I am still a grateful student, learning how to make our partnership even stronger.

Hema, thank you for being the constant in this incredible ride. I am deeply grateful for everything we have built together, and excited for all that still lies ahead.

❤️ Happy 25th wedding anniversary, my dear. ❤️

ServiceNow Claude

ServiceNow is officially making Anthropic’s Claude the heartbeat of its platform. 

This isn’t just a tech upgrade; it’s an AI takeover. By deploying autonomous agents for clients like Panasonic Avionics, ServiceNow is moving beyond traditional SaaS and becoming the AI workflow backbone for complex, regulated industries.

For the market, the message is loud and clear: ServiceNow is no longer just managing your workflows—it’s automating them at scale

Press release at https://newsroom.servicenow.com/press-releases/details/2026/ServiceNow-and-Anthropic-partner-to-help-customers-build-AI-powered-applications-accelerate-time-to-value-and-apply-trusted-AI-to-critical-industries/default.aspx 

PS: On listening AI paradox recently at industry, proud to be part of my firm, to invest & inspire AI consistently in people/culture for last 14+ months.

AI Paradox

For years, enterprise software was the "safe bet" of Wall Street—predictable, subscription-based, and essential. But today, a "double-whammy" of disappointing outlooks from industry titans has sent the sector into a tailspin, as investors begin to fear that artificial intelligence isn't just a tool for these companies, but a potential replacement for them.

The tremors started with SAP, which saw its shares plunge more than 16%—its biggest daily loss in years—after issuing a 2026 revenue forecast that fell short of analyst expectations. The German giant’s cautious cloud outlook served as a warning light for the entire industry.

Closely following was ServiceNow, which dropped 11%. Despite reporting subscription revenue that actually beat estimates, the results weren't "convincing enough" to drown out the growing narrative that AI disruption might eventually shrink the market for traditional workflow tools.

For the first time in a decade, the "Software as a Service" (SaaS) model is being questioned. As AI agents and autonomous coding tools become more capable, the traditional software giants are racing to reinvent themselves through massive M&A deals and new AI-integrated features. 

20 AI daily tools

 

 Stop Working Harder, Start Working Smarter

We’ve officially moved past the "AI is coming" phase—AI is here, and it’s transforming how we handle everything from research to coding. If you feel like your to-do list is growing faster than you can keep up, it might be time to audit your toolkit.

The infographic below highlights 20 power players in the AI space to shift your workflow from manual to autopilot.

LLM steps

The rapid evolution of Large Language Models (LLMs) has transformed how we interact with technology. From writing code to offering medical insights, these systems feel almost human. But behind the "magic" lies a rigorous, multi-stage lifecycle. 

Based on the expert roadmap provided by Rocky Bhatia, here is a breakdown of how an LLM is built, refined, and maintained.

Stage 1: The Foundation – Data & Architecture

Before a model can "think," it must be fed. This initial phase focuses on gathering the raw materials of intelligence.

Stage 2: Training – Learning to Predict

With the architecture set, the model begins its "education."

Stage 3: Alignment – Fine-Tuning for Quality

A model that knows everything is useless if it isn't helpful or safe. This stage "aligns" the model with human values.

Stage 4: Deployment & Interaction

Once the model is accurate and safe, it is ready for the real world.

Stage 5: Evolution – Performance & Continuous Learning

The journey doesn't end at deployment. Language and information are always changing.

Enterprise AI strategy

 

This image perfectly captures the "AI Paradox" that many enterprise leaders are currently facing. While the board is excited about the shiny "AI Strategy" peak, the engineering teams are deep underwater dealing with the messy reality of technical debt.

Every executive team wants to "move fast" on AI. They see the summit—generative models, automated insights, and transformative growth. But beneath the surface of a sleek AI Strategy lies a massive, cold reality: the technical debt of the last twenty years.

If you try to build a high-speed AI engine on a foundation of "spaghetti" code and siloed data, the iceberg won't just slow you down—it will sink the project.

To move from the "Excited Executive" boat to the "Successful Implementation" submarine, teams need to stop treating AI as a separate project and start treating it as the ultimate reason to clean house.

• Audit Before You Automate: Document your legacy logic before asking an AI to refactor it.

• Data Hygiene is AI Strategy: If your data is messy, your AI will just be "fast and wrong."

• Empower the "Foundations" Team: Give the engineers in the middle boat the resources they need to fix the base. You can't reach the peak if the bottom is crumbling.

2026 Case Studies

Theoretical knowledge of databases and load balancers is a start but seeing how they are applied under extreme pressure is where the real learning happens. These case studies teach you about:

• Scalability: How to grow from 1,000 to 1 billion users.

• Availability: Ensuring services never go down, even during updates.

• Latency: Reducing the milliseconds between a user's click and the system's response.

1. YouTube & MySQL (The Power of Sharding)

Most people assume a giant like YouTube would abandon traditional relational databases for something "fancier." However, YouTube famously scaled MySQL by using Vitess, a database clustering system.

• The Lesson: You don't always need a NoSQL database to scale; horizontal sharding can allow traditional SQL databases to handle massive throughput.

2. Uber’s Geospatial Indexing

Uber processes over 1 million requests per second to match riders with drivers. This isn't just a simple query; it requires calculating distances in real-time.

• The Lesson: Uber uses hexagonal hierarchical spatial indexing (H3). By dividing the world into hexagons, they can quickly calculate "proximity" without taxing the CPU with complex geometry.

3. Meta’s Serverless Scale

Handling 11.5 million serverless function calls per second is an exercise in extreme orchestration. Meta (Facebook) has to manage resource allocation and "cold starts" at a scale most companies will never see.

• The Lesson: Serverless isn't just for small hobby projects; with the right infrastructure, it can power the world's largest social network.

System design is less about finding the "perfect" tool and more about understanding trade-offs. Whether it's choosing between the consistency of Google Docs or the massive throughput of Kafka, these 12 examples provide the blueprint for modern engineering.