I build AI products from zero to one, and I teach others how to do the same.Over 12+ years, I've taken AI/ML products from proof-of-concept to launch at Amazon, Nike, T-Mobile, and Disney. The work I love most is early-stage: validating ideas, getting executive greenlight, and shipping MVPs that prove value before scaling. At Nike, that approach drove a $100M annual reduction in product returns.I also teach what I build. I co-created and lead the AI Product Management Certificate at the University of Washington, where experienced PMs ship real GenAI products complete with evaluation plans and GTM strategies. Previously, I trained 4 cohorts at Product School.What I bring: the ability to translate messy AI capabilities into clear product value, align technical and business stakeholders, and move fast without breaking trust. I write about AI PM craft at aipmguru.substack.com.I'm currently exploring opportunities to lead 0→1 AI product work at companies solving meaningful problems. I'm especially energized by AI in healthcare and education, though I'm open to any domain where the product can genuinely improve lives. 

Shaili Guru

Hemant Bothra and I are running an in-person workshop where we're going to build something together - in the same room, like the old days.

3 hours. You'll leave with a working tool.

No coding experience needed (you read that right). We'll use Cursor (an AI coding assistant) and walk through everything step by step. If you've ever been curious about "vibe coding" but didn't know where to start, this is it.

Good fit if you're a PM who wants to actually build things, or just someone who learns better by doing than watching.

AI/ML Product Leader | Driving Innovation with AI | PM Mentor and Advisor

Open article: Build Your First AI Tool: ChatGPT Output Cleaner Workshop · Luma by luma.com

Build Your First AI Tool: ChatGPT Output Cleaner Workshop · Luma

AI Product Builder @ NanoKernel | Generative AI, AI Agents, AIoT, Responsible AI, AI Product Management | Ex-Apple, Accenture, Cognizant, Verizon, AT&T | I help companies build standout Next-Gen AI Solutions

I help visionary companies build standout AI Products. I am seeking a Senior IC AI Product Manager role at established AI-focused companies. I am a transformative Product and IT Executive, blending innovation with leadership across every role I've embraced. My journey from the technical trenches to the strategic echelons of Product Leadership, Solutions Architecture, and Enterprise Data Strategy has been fueled by an insatiable drive for growth, an unwavering focus on customer success, and a passion for adapting and thriving in diverse organizational landscapes.From Fortune 50 giants to agile startups, my path has been marked by over 8 groundbreaking product launches and more than 20 transformative project deliveries. At the heart of these achievements lies a deep-seated belief in the power of innovative thinking, strategic execution, and the magic of collaboration across domains and disciplines. My mantra? Value creation, operational excellence, and nurturing talent are the cornerstones of my transformative leadership.My commitment to thought leadership has led me to author a book on Big Data, secure several patents, and contribute to the knowledge pool with articles and papers featured in renowned journals, blogs, and media channels. Dive into my insights on LinkedIn and Medium at https://harsha-srivatsa.medium.com/Education is an ongoing journey for me, enriched by formal degrees in Engineering, Math, Computer Science, and Management, alongside continuous learning from prestigious institutions like MIT, Stanford, Kellogg, and IDEOU. This foundation empowers me to lead at the intersection of technology and business, navigating the complexities of AI, IoT, Cloud Platforms, and beyond with ease and insight.My commitment extends beyond the professional realm. As a High School Educator with Microsoft's TEALS Program, a Master Gardener certified by University of California, a mentor for immigrants through Upwardly Global, and a volunteer with the Coast Guard Auxiliary, I believe in giving back, sharing knowledge, and supporting communities.I am a believer in the transformative power of technology, education, and community service. Let’s connect, collaborate, and create impact together.Email:  harsha.srivatsa at connectwithharsha.info | Website: https://www.connectwithharsha.me

Harsha Srivatsa

Thanksgiving 2025…

For me, it’s not about the food, the shopping, or the holiday itself. It’s about the occasion—and the sentiment of pausing to say thank you.
This Thanksgiving, I’m giving thanks for the people who have touched my life, the places that have shaped me, and the possibilities that make life purposeful.

Thank you to two amazing people I met recently—Polly M Allen and Nico Posner —who filled my heart and soul with joy.

Thank you to the beautiful Bay Area that my wife and I have called home for 2+ decades.

Thank you to SFO, my City of Possibilities, where I discovered—and continue to build—my passion for creating AI products.

Thanks for being part of my life journey and for reminding me, in big and small ways, how rich my life becomes when I experience gratitude and share it.

#ThanksgivingGratitude #PeoplePlacesPossibilities #AIPM

Senior AI Product Leader | Generative AI, AI Agents, AIoT, Responsible AI, AI Product Strategy | Ex-Apple, Accenture, Cognizant | I help companies build, debug and launch standout AI products

I am continually fascinated by the recent LLM Council project by Andrej Karpathy. I now consider the LLM Council as both a Gen AI Design Pattern and an application building block. The possibilities for building innovative applications seems to be quite a lot. 

The LLM Council pattern embodies the principle that N models > 1 model when properly orchestrated. Instead of relying on a single AI to answer your question—with all its biases, blind spots, and hallucination risks—what if you assembled a "council" of diverse AI models to debate, critique, and synthesize a superior answer?

Today, as I was working on a PRD, I had the Aha! moment to explore the application of LLM Council Pattern to PRD reviews and rapid iteration and improvement. 

We all know this: the real gaps in a PRD show up after sprint planning – during engineering review, in QA, or (worst) in production.

I came up with an application concept: PRD Review Council – an AI “board of PM advisors” powered by the LLM Council pattern that reviews my PRD before it ever reaches sprint planning.

Here’s the idea 👇
Instead of a single AI doing a generic review, my PRD (generated the classical way or via ChatPRD)  goes to a council of AI personas inspired by product thought leaders – think Jackie Bavaro (clarity & strategy), Melissa Perri (outcomes over output), Dan Olsen (PMF & lean experiments), Janna Bastow (roadmaps & process), Claire Vo Vo (AI-native PM & agents). 
Each one reviews the PRD from their own lens and surfaces issues, questions, and suggestions. 

Then a “chairman” persona inspired by Marty Cagan pulls it all together: synthesising the feedback into one coherent review with: 
* An executive summary of readiness
* Critical gaps that block development
* High-leverage improvements to problem, outcomes, scope and risks
* A short list of concrete next steps for the author PM

Under the hood, PRD Review Council uses an LLM Council-style workflow: multiple models/experts in parallel, cross-review of each other’s feedback, and a final synthesis step that optimises for completeness, coherence, and practical actionability – not just clever comments. 

Why this matters:
* Catch most structural gaps before sprint planning
* Reduce back-and-forth PRD revisions with engineering, design and leadership
* Turn every PRD review into a mini coaching session from multiple “schools of PM”
* Build a repeatable, high-quality PRD review ritual even in lean teams 

I’m exploring this both as a real product idea implemented with Google AI Studio and as a reference app for multi-LLM “council” architectures. 

If you’re a PM or product leader who struggles with PRDs that get shredded late in the process, I’d love your take:

👉 Would you trust (and use) an AI PRD Review Council like this? What would you want it to catch every single time?

#ProductManagement #AIProductManagement #LLMCouncil #GenerativeAI #ProdMgmt

We now have a menu of LLM platforms and frontier models. 

OpenAI, Google DeepMind, Anthropic, Perplexity, xAI, Mistral AI, DeepSeek
and others keep shipping new model releases, each with fresh claims about features and performance.

But a practical question remains: which model should I use for which task, purpose, and outcome? How do we look beyond leaderboards and hype, and relate these models to day-to-day, pragmatic work?

I found comparing frontier LLM platforms like ChatGPT 5.1, Claude 4.5, Grok, Perplexity, and Gemini quite difficult because they are complex products, not just models on a chart. I needed a structured way to map their distinct strengths, risks, and task fit so I could confidently route work to the right tool.

So I adapted Andrej Karpathy’s “LLM Council” idea — a multi-model debate plus chairman synthesis — to systematically evaluate each LLM platform.

The result is a practical view, for each platform, of:
* Strengths
* Watch-outs
* Tasks well suited
* Tasks better routed elsewhere

The adapted LLM Council process:
-> Multiple frontier models independently assess a target platform to avoid early anchoring.
-> They anonymously critique and rank each other’s answers to surface errors and nuances.
-> A “Chairman” model, overseen by a human, synthesizes the debate into a consensus verdict.

The Council’s one-sentence report snapshots:
> ChatGPT 5.1 – A top-tier generalist with adaptive reasoning that excels at explaining concepts, agentic workflows, and acting as a conversational copilot.
> Anthropic Claude 4.5 – The “deep thinker” for complex coding, heavy document analysis, and long-horizon reasoning, especially when paired with Artifacts.
> Perplexity – A web-native “answer engine” built for deep research and fact-finding, rather than long-form creative writing or storytelling.
> Grok – Best for real-time social trend analysis and sentiment tracking, with a spicy tone and direct access to X data.
> Google Gemini – A multimodal powerhouse that integrates with Workspace and uses massive context windows to analyze video, large codebases, and Drive files.

Comment “Council” if you’d like the full breakdown and report.

The attached slide deck captures and summarizes the key findings.

Doing this exercise gave me a much clearer mental model of which LLM platform to use for what, and how to route tasks to the best-fit model for the outcomes I care about.

#AI #GenerativeAI #AIProductManagement #FrontierLLMs #AIEvaluation #AISafety #LLMCouncil #ProductManagement

Evaluating LLM Platforms with the LLM Councils

Sharing a personal spirit lifting milestone from ResearchGate. 

Glad to know that people are reading my published paper. Inspired by my mentor Aman Chadha and by his stellar publishing record and portfolio. 

This path to publishing preprints and getting reads / citations is to get into the groove of original ideation, thinking about problems - solutions, and putting high quality written work. This effort with ResearchGate is a start; I aim to work and uplevel myself to be worthy of acceptance into "prestigious" journals - arXiv, IEEE, ACM etc. It is hard but I am motivated to do whatever it takes to get to that level.

🚀 Book Release Announcement - Nov 21 2025! 

I’m excited to share that my book AI Product Builder’s Troubleshooting & Diagnostic Guide is now complete and in final refinement. It will be released on Nov 21, 2025 and sent directly to everyone who signed up. 🙏

If you’d like to receive the e-copy, you can still join the list here:  👉 https://lnkd.in/gz2mVPe6

Earlier announcement: https://lnkd.in/ga5Qd9vy

💡 What makes this book different
To make this a practical, usable manual for real-world AI teams, I tried several first-of-its-kind approaches:
• Each chapter is exactly one page — sharp, concise, and built for busy AI builders

• TL;DR + Bottom Line — know the “why” and “what to do” instantly

• A unified visual synthesis of the entire book

• Multimodal chapter experiences:
  🎥 ~8-min video explainers
  🎧 Dialog-style audio versions
  📘 Detailed text + visuals
 
Designed to support every learning style.

🛠 Why I wrote this
AI Product Builders don’t need more “how to build a RAG app” tutorials.
We need guidance on:
 ✔ debugging failures
 ✔ diagnosing drift
 ✔ preventing cost explosions
 ✔ hardening AI for real production
 ✔ governing responsibly at scale

A mentor summarized it perfectly:
“This isn’t just a book — it’s an intervention… You’ve addressed the ‘Pilot Purgatory’ problem every VP of AI feels right now.”


📘 What this book aspires to be
A living operational manual that unifies Product + Engineering + AI implementation realities.

A playbook to help teams:
 • cross the pilot-to-production gap
 • debug with clarity
 • apply diagnostic hardening
 • evaluate, govern, and scale with discipline

After launch, I’ll also release templates, guides, Miro boards, and walkthrough videos to make the playbook even more actionable.

This work is dedicated to every AI Product Builder striving to build their “Aha!” product — responsibly, reliably, and at scale.

#AIProductManagement #AIProductBuilders #AILifecycle

AI Product Lead @ NanoKernel | Generative AI, AI Agents, AIoT, Responsible AI, AI Product Management | Ex-Apple, Accenture, Cognizant, Verizon, AT&T | I help companies build standout Next-Gen AI Solutions

I am thrilled to announce a collaboration with my friend Sevanne Calle on upleveling ourselves in Physical AI! 

To start off, we have created a newsletter titled "Physical AI Dispatch" 

Issue #1 is now live: “Why Physical AI Matters Now.”
In this first issue, we break down:
* What I mean by Physical AI (beyond the buzzwords)
* How it will reshape work, industries, and everyday environments
* Why now is the time for AI Product builders, AI leaders, and AI practitioners to upskill and get ready

If you’re curious about the intersection of AI + Robotics + AIoT + Edge Infrastructure, or you’re an AI builder, product leader, engineer, or operator who wants to ride the Physical AI wave,  I’d love for you to follow along.

🔔 Subscribe to Physical AI Dispatch to get future issues, deep dives, and practical frameworks straight in your inbox.
[https://lnkd.in/gubSgEYB]

We’re entering a moment where AI is no longer staying in the digital world.
It’s stepping into the real world — into devices that sense the environment, robots that move with purpose, machines that can help us in hospitals, farms, factories, and even in our homes. This shift can change industries, careers, and lives.

But right now, most people only understand “software AI”: LLMs, chatbots, copilots.

Physical AI is the next wave — and most people are not prepared for it.

Over the past months, I’ve become convinced that Physical AI — AI embedded in robots, machines, vehicles, and devices that can sense, decide, and act in the real world — will define the next decade of innovation. From logistics and manufacturing to healthcare, energy, agriculture, and smart cities, we’re already seeing early signs: autonomous robots, intelligent fleets, warehouse automation, hospital logistics, and more.

I want to prepare myself now for that shift — to think deeply about:
* How we design Physical AI systems that are helpful rather than harmful
* How we blend AI, robotics, AIoT, and edge infrastructure responsibly
* How Product, Engineering, Operations, and Policy need to work together to make Physical AI safe, reliable, and truly human-centered

👉 So please join us in the journey of this exciting new world of possibilities and potential from Physical AI. 

#PhysicalAI #Robotics #AIoT #AIProduct #FutureOfAI

I have been using the new Slide Deck generation feature in Google NotebookLM and its pretty awesome - provided you give it some clear and good instructions. 

Here is a sample slide deck generated for Chapter 2 - The Pilot-to-Production Bridge in my book The  AI Product Builders Troubleshooting Guide.

The AI Product Builder's Guide - Chapter 2 Slide Deck

🚀 It’s here! I’m excited to officially release my new book:
 “AI Product Builder’s Troubleshooting & Diagnostic Guide”

Today, I’m releasing Version 1 - the first edition of what I hope becomes a living, evolving manual for AI Product Builders everywhere.

🔗 Download / Access the book:
Formatted PDF (recommended for visuals):
https://lnkd.in/gQAqvWNY

Living Google Doc (always up-to-date):
https://lnkd.in/gguqQTEH

📝 Suggest updates / fixes / improvements:
 👉 Feedback Sheet: https://lnkd.in/g7x622X8

💡 Why I wrote this
AI Products don’t fail at the model stage—they fail at the system stage. Drift, bad data, fragile pipelines, unclear governance, cost explosions, missing guardrails — these are all diagnosable and fixable patterns.

This book is the operational guide I wish my AI team(s) had:
On How to debug, diagnose, harden, evaluate, govern, and scale AI systems that must survive real-world conditions.

🚀 What makes this book unique: 
I designed this book with several first-of-its-kind features:
✨ Each chapter is exactly one page
 Clear. Sharp. Built for busy AI Builders.

✨ TL;DR + Bottom Line on every page
 No fluff. Immediate clarity.

✨ A unified visual synthesis in each chapter. 
Each chapter content synthesized into one artifact.

✨ Integrated multimodal learning
 For every chapter, you can choose to:
 🎥 Watch an ~8-min explainer video
 🎧 Listen to a dialog-style audio version
 📘 Read the text + visuals

✨ Living manual — continuously updated
 The Google Doc will remain the source of truth.
 I will track updates in a public change log and release refreshed versions frequently.

🛠 How you can use the book
Use it as your:
 • Production hardening checklist
 • Pilot-to-Production playbook
 • Evaluation & guardrail reference
 • Incident response guide
 • Governance + FinOps alignment tool
 • Workshop toolkit for AI maturity

My goal: help teams escape “Pilot Purgatory” and build reliable, safe, cost-aware, trustworthy AI systems.

A mentor who read this book earlier, captured it well:
“This book isn’t just guidance — it’s an intervention. It tackles the exact pain point every VP of AI is wrestling with right now.”

🌱 Your feedback matters
This is Version 1 — the beginning.
 Please use the feedback sheet to propose:
 • Fixes
 • Improvements
 • Clarifications
 • New chapter ideas
 • Real-world cases
I’ll review every submission and integrate improvements continuously.

🙏 Final note
I’d be grateful if you could:
 💬 Share this post
 📤 Forward the book
 🏷️ Mention it to teams who need it
 📚 Use it in your org’s AI processes

Thank you for your support — this book is dedicated to every AI Product Builder striving to build their “Aha!” product responsibly and at scale.

#AIProductManagement #AIProductBuilders #AILifecycle

What happens when AI’s hunger for power, water, and land collides with the limits of our planet – and the patience of our communities?

We’re already seeing the early signs: grid constraints, water stress, local moratoriums on new data centers, and growing scrutiny from regulators and citizens. 
At the same time, we’re racing to build an AI++ world – where AI is deeply intertwined with IoT, edge, robotics, and cyber-physical systems.

To make that future possible, we don’t just need more  power or more AI infrastructure.

We need to have a clear vision for the Future of Data Centers, And We need Resource-Responsible AI so that we invest and build in that AI infrastructure rationally. 

I’ve pulled together my research, thoughts, and field notes into a publication: 
📘 “The Future of AI Data Centers – Creating Resource-Responsible AI for an AI++ World”

Links to the booklet and slide deck in the comments. 

In this publication I explore:
* How AI data centers differ fundamentally from traditional cloud infrastructure
* Why power, water, and cooling are becoming design constraints – not afterthoughts
* A set of principles and scorecards for building Resource-Responsible AI across the model lifecycle
* Patterns for AI-native architectures spanning edge, micro data centers, and hyperscale
* The role of regulation, risk, and “social license to operate”
* Future scenarios and a practical roadmap for teams who want to act, not just admire the problem

If this topic resonates with you – whether you’re in AI/ML, cloud, data centers, sustainability, or policy – drop a comment or DM  🙌

#AIDatacenters #ResourceResponsibleAI #SustainableAI #AIInfrastructure #AIoT #EdgeComputing #AIArchitecture #AIProductLeadership

An update to my earlier post announcing the upcoming release of my book titled "AI Product Builder's Troubleshooting Guide" 
https://lnkd.in/ga5Qd9vy

Today I want to introduce a key concept in the book - Seven Bridges. 
I am so much fascinated by the term. 
Seven Bridges of Konigsberg is a classical mathematics / computer science problem. 
Seven Bridges Road is a beautiful song by The Eagles. 

In AI Product Building, The Seven Bridges diagram maps the chasm and the critical transition gaps between pilot success and production scale. It connotes the difference whether an AI project is destined to Pilot purgatory or whether it makes it a successful Production GTM. It is a diagnostic model identifying the precise failure points where AI projects fracture during scale-up. Each bridge—Data Readiness, Model Robustness, Infrastructure Scalability, Operational Maturity, Governance Readiness, Organizational Alignment, and Business Value—represents a critical handoff where assumptions break, ownership blurs, and risk compounds. Looking at an AI Project from the Seven Bridges lens can help cross the chasm to a successful AI project. 

For this book, my content creation approach has been: 
- Generate a text document with concepts, principles, visual artifacts, practices etc. 
- For each chapter, I have created a companion video explainer to amplify and explain the contents in 8 mins or less
- And an audio overview with two hosts discussing and debating the insights and nuances. 

In short, a multimodal approach to convey what I want to share. 

As an example, here is the content for Chapter 1: 
Text: https://lnkd.in/g6hdBiWn
Video Explainer: https://lnkd.in/gaEhUa2u
Audio overview: https://lnkd.in/gYXEt-j5 

I hope you like to content creation approach and formats. Let me know in comments. 

#AIProductManagement #AIProductBuilder #AIProductDevelopment

Thrilled to share that my new research paper is now available on ResearchGate: “Agentic Swarm Intelligence Applied to AIoT Agent Coordination: A Domain-Agnostic Framework for Multi-Agent Coordination, Task Allocation, and Human-AIoT Workflow Delineation.”

This work brings together Swarm Intelligence, multi-agent systems, and AIoT to answer a key question: 
How do we coordinate AIoT agents while keeping humans safely and meaningfully in the loop? 

The paper introduces the AIoT Swarm Coordination Map (ASCM), an extended RACI-A responsibility model, and a Human–AIoT Workflow Canvas that practitioners can apply across manufacturing, energy, smart cities, logistics, and more.

If you’re interested in and exploring Multi-agent systems, AIoT, edge intelligence, or Human-AI teaming, I’d really value your read, critique, and suggestions for future directions.


#AIoT #AgenticAI #SwarmIntelligence #MultiAgentSystems #EdgeComputing #HumanInTheLoop #AIArchitecture #EdgeAI

Going From Vibe Synthetic Multimodal Data Creator to a Vibe  Synthetic Reality Creator! 

Today I got access to https://lnkd.in/gQvPJ5Kz World Labs
World Labs (the startup founded by Dr Fe-Fei Li https://lnkd.in/gMiebVzw Fei-Fei Li
Marble 3D World creation platform. And I am blown away by the possibilities. 

If I can think / imagine a scenario, I can now create  Synthetic Reality Worlds. The possibilities for AI Products with Visual Intelligence, Robotics training, and simulation-first design feel enormous.

With Marble + Veo3, it’s becoming possible to go beyond  synthetic multimodal images and videos… and instead construct entire 3D, physics-aware simulation reality worlds. 

This has so much potential for AIoT systems, robotics, and digital twins.

This is a game changer. This has so much value for my work. 

Imagine designing scenarios that are hard, dangerous, or impossible to capture in the real world:
 🔥 fire and explosion events
 🪖 battlefield and high-risk environments
 🤖 humanoid robot training scenes
 🏥 hospital digital twins
 🌆 smart-city anomaly simulations

Instead of creating the data, we create the reality with the data! 

My first experiment was to use WorldLabs Marble to generate a 3D panorama of a battlefield scenario as it would appear for an autonomous drone as it surveys the battlefield scene for sitrep and damage assessment. I then used to panorama snapshots to create a realistic battlefield scenario video. 

Early results are promising. See the generated video attached and the panorma screenshots in the comments. 

If you’re building in digital twins, simulation, robotics, or AIoT, keep an eye on this space.

Synthetic Reality is coming fast.

#AI #SyntheticData #Simulation #AIProductManagement #AIoT #Robotics #DigitalTwin #AgenticAI #Marble #Veo3

A TGIF moment! Thankful to meeting and collaborating with awesome people like Ronny Ashar .

It is an awesome feeling for me when you can have meaningful connects even though we are thousands of miles apart and we can find something that we are both passionate about. 

Ronny and I have had deep and thoughtful discussions about Sustainable AI recently. Which led me to the formulation of the concept of Resource Responsible AI. 

I am really proud of Ronny for getting the AI Ethics practioner certification from Stanford. Looking forward to continuing this awesomeness in 2026.

taplio

The Battle

Shaili Guru

Harsha Srivatsa

Followers

Posts

Engagements

Likes

Performance ScoreShaili Guru's LinkedIn Performance

Performance ScoreHarsha Srivatsa's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Shaili Guru

Harsha Srivatsa

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Shaili Guru's LinkedIn Performance

Performance ScoreHarsha Srivatsa's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreShaili Guru's LinkedIn Performance