Agentic AI Execution

Could your program survive two weeks without you? If the answer is no, you’re not running an operation—you’re a single point of failure with a job title.In this episode of Agentic AI Execution, host Leena sits down with Ankur Gupta, Senior Staff TPM leading AI Infrastructure at LinkedIn (and former tech leader at Walmart). They tackle the burning question: Will AI replace technical program managers?Ankur breaks down why AI isn't taking your job, why status reporting was never the core value of a TPM, and how to build a "TPM Operating System" that runs seamlessly with AI agents while keeping human judgment at the center.---📌 KEY TAKEAWAYS & HIGHLIGHTS:• The Two-Week Test: How to structure programs that don't depend on your memory.• Status vs. Judgment: Give agents the status updates, but keep judgment to yourself.• Information Centralization & Work Decentralization: Scaling execution across 200+ teams.• The TPM Operating System: Operational, Collaboration, and Technical lenses.• Guardrails Over Memos: How to build real AI governance directly into technical systems.• Deterministic vs. Non-Deterministic Reliability: Managing AI agents that take different paths to a goal.---⏱️ TIMESTAMPS:00:00 - The Two-Week Test: Are You a Single Point of Failure?01:03 - Ankur Gupta's Career Journey to LinkedIn AI Infrastructure02:20 - What Agentic AI Execution Means in Practice03:30 - The Evolving Role of a TPM in the AI Era06:05 - Information Centralization vs. Work Decentralization08:45 - Moving from Coordination to Strategic Tech Leadership11:20 - The "TPM Operating System" Framework17:12 - Live Walkthrough: Set Once, Run Forever Checklist22:50 - What Should NEVER Be Handed Off to an AI Agent?25:10 - Lowering the "Cost of Yes" & Influencing Humans28:30 - Who Holds Accountability When AI Agents Ship Code?30:50 - AI Governance & System Guardrails34:10 - Reliability in Non-Deterministic AI Systems37:40 - What TPMs Need to Do in the Next 12 Months to Stay Relevant42:15 - Practical Hands-On AI Hacks for Daily Workflows46:10 - Overcoming AI Anxiety & Final Advice49:18 - Key Takeaways & Episode Wrap Up🔗 Connect on LinkedIn: Ankur : https://www.linkedin.com/in/agupta03/Leena: https://www.linkedin.com/in/leenaagarwal/

In this episode, I sit down with Michael—a technical product management leader who has spent the last ten years building TPM teams from scratch. Together, we explore what agentic architecture looks like inside an engineering org and the mentorship that separates teams winning with AI from those who are just experimenting. Michael shares his mental models for AI adoption, explaining why we need to move past the "experimentation phase" and focus on building durable systems. Whether you are a product manager, TPM, or engineer, this conversation provides practical advice on shifting your perspective from using AI as a tool to leading AI as a teammate. ——————————————————————⏱ CHAPTERS——————————————————————00:00 - Introduction to Agentic AI Execution 00:31 - Michael's Journey & Building a TPM Team 06:07 - Why 95% of AI Agents Fail to Reach Production 15:51 - How Engineering Teams Can Close the AI Adoption Gap 12:54 - Leading AI as a Teammate, Not a Tool 23:19 - Outro & TPM Breakdowns Substack Links & ResourcesMichael's substack: https://michigoetz.substack.com/aboutConnect with Us!

• Michael: https://www.linkedin.com/in/michi-goetz
• Leena:https://www.linkedin.com/in/leenaagarwal

In Episode #17 of the Agentic AI Execution Podcast, host Leena sits down with Myra Deshmukh—a 10-year Amazon veteran and founder of Leadership Lab—to break down what leadership actually looks like when half your team consists of AI agents. If you've been feeling "Agentic AI FOMO" or overwhelmed by how fast tech is moving, this episode is a must-watch. Myra shares her golden rule for AI, how to avoid outsourcing critical thinking, and why human judgment is your greatest asset in an automated world ---------------------------------------------------⏱️ TIMESTAMPS---------------------------------------------------00:00 - Intro: Leading when half your team is AI agents 00:32 - Welcome Myra Deshmukh (Leadership Lab & Ex-Amazon) 01:15 - Dealing with uncertainty & Agentic AI FOMO03:40 - Delegate work, keep judgment: Avoiding over-validation09:30 - How leadership style is changing in top tech companies 14:10 - Simple ways to build authentic relationships & trust at work 21:15 - The Manager's Insurance Policy: Verifying agent output29:50 - What to do if you feel overwhelmed or behind with AI 38:10 - Final advice for leaders & wrapping up ---------------------------------------------------🔗 RESOURCES & LINKS---------------------------------------------------• Check out Leadership Lab: https://www.leadership-lab.co/• Connect with Myra Deshmukh: https://www.linkedin.com/in/myra-deshmukh/• Leena | Agentic AI Execution Podcast https://www.linkedin.com/in/leenaagarwal/ #AgenticAI #Leadership #ArtificialIntelligence #FutureOfWork #ProductManagement #LeadershipLab #TechLeadership

With AI accelerating data flow, human alignment and judgment have become the ultimate bottlenecks in product development. If you are still relying on traditional, slow decision hierarchies, you are falling behind in the age of agentic intelligence.In this episode of *Agentic AI Execution*, Anjali—who leads AI programs on Google Maps and mentors over 100 startup founders—breaks down how to deliver AI products at a scale of billions. We explore why human-in-the-loop is still non-negotiable despite rising automation, how the Technical Program Manager (TPM) role is morphing into a multi-hyphenate AI builder, and the practical agentic workflows she uses every day to reclaim her time .*⏳ Timestamps:*

• *0:00* - Introduction: Leading AI at Google Maps and mentoring 100+ AI founders
• *0:27* - Product thinking for 1 user vs. 1 billion users
• *1:10* - How LLMs changed data analysis and product architecture
• *2:23* - Human alignment is the new bottleneck in tech
• *3:30* - The shift from 5 decisions a week to 20 decisions a day
• *3:37* - Why Subject Matter Experts (SMEs) are taking over leadership decisions
• *4:00* - Automation vs. Human-in-the-Loop: An AI reporting case study
• *5:55* - Why AI is a database, not a thinker
• *6:23* - Will TPM, PM, and Engineering roles merge into one?
• *7:45* - How AI makes product leaders proactive instead of reactive
• *9:33* - How to stay relevant and "add value + 1" in an AI-driven market
• *11:00* - Transitioning careers: From Admin or UX Designer to PM using AI
• *12:00* - Anjali's AI stack: Gemini, NotebookLM, Claude, and Anti-Gravity
• *13:50* - Predictions: Agents will own everything on Google Maps in 2 years
• *15:00* - How Anjali uses daily AI agents to filter emails, pings, and tasks
• *16:23* - Final thoughts: "No one can take your job but you"

• *Anjali:* [  / anjalikakkadp  ](  / anjalikakkadp  )
• *Leena:* [  / leenaagarwal  ](  / leenaagarwal  )

*🔗 Connections:*

**The internet has fundamentally changed.** Over 60% of web traffic is now driven by AI agents, not humans. If you are still building landing pages and products exclusively for people, you are missing the biggest shift since the dot-com boom.In this episode of *Agentic AI Execution*, Mariane—Head of DevRel at You.com and founder of the Founders Bay community—breaks down why we are entering an AI renaissance rather than a bubble. We discuss why developers must evolve into "full-stack AI engineers" who manage 10-20 agents simultaneously to stay competitive. We also dive deep into the reality of building a massive community from scratch, the failure patterns of early-stage founders building in isolation, and the very real balancing act of scaling a career while managing mom guilt.**⏳ Timestamps:**0:00 - Over 60% of web traffic is now AI agents0:37 - Mariane's journey: Engineering leadership to You.com DevRel2:21 - The #1 reason AI founders fail: Building in isolation4:47 - Why you need to build websites for agents, not humans6:13 - AI Bubble vs. The AI Renaissance7:45 - Vibe Coding: Managing 10-20 agents at once9:11 - Handling bugs and CI/CD with AI-generated code11:21 - The rise of the Forward Deployed & Full Stack AI Engineer14:50 - Growing Founders Bay and building community from scratch18:30 - How to build an audience from zero today22:15 - The future of DevRel in an Agentic AI world23:30 - Managing burnout, parenting, and the reality of "mom guilt"26:06 - Closing thoughts: Keep building and showing up🔗 Connections:* Mariane: https://www.linkedin.com/in/marianebekker/ * Leena: https://www.linkedin.com/in/leenaagarwal/* Founders Bay Community: https://www.linkedin.com/company/foundersbaynetwork

What if the hardest problem in AI agents isn't memory — it's judgment?In this episode of Agentic AI Execution, I sit down with Adelina — former colleague at Cruise, later at Vercel, and now founder of Rory Plans, building AI agents for organizational decision-making.We get into why agents with perfect memory still make the wrong recommendation, why bad decisions compound faster than you'd think, and what she wishes she'd known before quitting her job to start her company.In this conversation:→ Why memory tells an agent what happened — but not what matters→ What Cruise taught her about decisions under uncertainty, and how it applies to AI agents today→ Why Rory Brain exists to capture judgment, not just information→ Running "Lunch and Hack" events and what builders are getting wrong with AI agent memory→ Why she believes organizational intelligence — not bigger models — will define the winners in AI→ Her advice to founders: spend 80% of your time with customers, 20% building🔗 Connections:* Adelina: https://www.linkedin.com/in/adelinamartiniuc/* Leena: https://www.linkedin.com/in/leenaagarwal/* Rory Plans: https://roryplans.aiTimestamps:00:00 Intro00:26 Adelina's journey: Cruise → Vercel → founding Rory Plans02:30 Why judgment is harder than memory for AI agents05:00 What Cruise taught her about decisions at scale07:30 Building Rory Brain to capture organizational judgment10:30 Lunch and Hack: what builders get wrong about agent memory13:30 Why organizational intelligence will define AI's winners17:00 Advice for founders: customers over building20:05 Closing takeaways + where to find Adelina#AgenticAI , #AIAgents , #TPM , #AI

*Are you vibe coding your way into a massive cloud billing nightmare?*AI agents are already writing code, reviewing ads, and making autonomous decisions. But who is actually watching them? In this episode, Leena sits down with Lulu, a data infrastructure expert who has scaled systems at Capital One, Block, and Reddit.They break down the "real talk" of scaling AI safely, why AI-generated code is breaking in production, and how a single unmonitored API key can cause a $20,000 disaster in less than 48 hours. If you are building with LLMs, this breakdown of infrastructure, guardrails, and cost-optimization will save your budget and your production systems.👇 *SUBSCRIBE* for more deep dives into the reality of engineering the AI boom.⏱️ Timestamps & Highlights

• *00:38* — Lulu’s journey: Inside data teams at Capital One, Block, & Reddit.
• *01:10* — Why your data infrastructure won't work at AI scale.
• *02:30* — How Reddit automates 90% of ad reviews using Gen AI enrichment.
• *04:30* — The hidden cost trap: Snowflake & Databricks vs. LLM scaling.
• *06:20* — Why running massive frontier models in production will "burn your car".
• *07:05* — Stop using frontier models for simple tasks (The Small Model Strategy).
• *09:20* — Why Stanford & Berkeley benchmarks are lying to you about real enterprise costs.
• *11:15* — What is "Contract-First Development" and why does AI code break?
• *15:00* — Building "seatbelts" for AI: Secure Kubernetes pods and real-time execution auditing.
• *18:15* — Startups vs. Enterprises: Compliance, SOC2, and the future of agent security.
• *21:30* — The Nightmare Scenario: Leaking API keys to your local coding agent.
• *23:00* — Lightning Round: The one hot take on AI code that engineers HATE.

🔗 🔗 Connections:

• Connect with Lulu:   / luluchengengineeratheart  
• Connect with Leena:   / leenaagarwal  

Engineering velocity just skyrocketed. Code is being written, tested, and shipped 10x faster than ever before thanks to generative and agentic AI.And yet — product failure rates are rising.Why? Because speeding up execution doesn't automatically mean progress. The bottleneck was never engineering. It was always judgment.In this episode, host Leena sits down with Revital — executive coach, startup advisor, and former product leader at Meta and Microsoft. With a career spanning BCG management consulting, finance, and chief of staff roles, Revital brings a deeply strategic lens to the AI transition every PM is navigating right now.We get into why Andrew Ng says PM is the new bottleneck in tech, how leaders use data to hide from hard decisions, and why human judgment under uncertainty is the ultimate competitive moat in the age of AI agents.If you're a PM, TPM, or tech leader navigating layoffs, AI anxiety, and shrinking team ratios — this is your blueprint.🔗 Work with Revital: invergency.com⏱️ Timestamps00:00 – Introduction: From BCG to Meta & Microsoft Product Leader01:25 – Navigating a Non-Linear Career (Letting Go of the Grand Plan)04:10 – What Shipping to Billions Teaches You About Psychological Safety07:35 – Safe Experimentation: B2B vs. B2C Startup Strategy11:05 – Is PM the New Bottleneck? The PM-to-Engineer Ratio Shift14:20 – Execution vs. Judgment: What AI Cannot Automate17:15 – The Decision Problem in Disguise (Are You Polishing Decks?)20:00 – The Human Edge: Why "The Glue" Matters More Than Ever22:50 – Imposter Syndrome, Perfectionism & Internal Narratives26:25 – Layoffs & AI Anxiety: The Truth About Losing Your Job to AI29:40 – Rebranding Networking: Connect Without Feeling Salesy31:10 – Rapid Fire: Self-Awareness & What a Great PM Looks Like in 5 Years

🔗 Connections:Connect with * Revital: https://www.linkedin.com/in/revitalpadovitz/* Connect with Leena : https://www.linkedin.com/in/leenaagarwal/

In this episode of Agentic AI Execution, Priyanka Shinde—founder of Solomate AI and former TPM leader at Meta and Cruise—joins the show. Following the massive wave of tech layoffs, Priyanka shares raw insights on why your job title might change, but your core execution and leadership skills are evergreen. We dive deep into how "English is the hottest new programming language", the reality of solo entrepreneurship, and how autonomous AI tools can buy your time back.

⏱️ Timestamps

• *00:00* – Intro: Is English the new programming language?
• *00:36* – From engineering back to building with cloud code
• *02:15* – Navigating the Big Tech layoff environment
• *04:10* – Why the TPM title is evolving into "AI Builders"
• *06:30* – Moving past chat prompts to autonomous AI
• *08:45* – Building Solomate AI for solo entrepreneurs
• *10:15* – Integration challenges: Context, judgment, and data privacy
• *13:20* – Can a one-person company hit a $1 Billion valuation?
• *15:45* – Strategic execution vs. leading with the solution
• *19:22* – Outro & 3 critical episode takeaways

🔗 Mentioned Links

• TPM Academy: https://www.tpmacademyai.com/
• Solomate AI: https://www.solomate.ai/

🔗 Connections: Connect with

• Priyanka Shinde:   https://www.linkedin.com/in/priyankash/
• Connect with Leena Agarwal:  https://www.linkedin.com/in/leenaagarwal/

Most companies fall in love with their AI on day one. Nobody asks what happens on day two.In this episode of Agentic AI Execution, I sit down with Noelle Russell — Employee #10 at Amazon Alexa, founder of the AI Leadership Institute, and closing in on 1 million views on her TED Talk.We cover:→ Why most AI systems go rogue (and who's really responsible)→ The principle of least privilege — and why leaders get it wrong→ What a "symphony of talent" actually means when building AI→ The baby tiger analogy that will change how you think about AI→ Why human-in-the-loop isn't a compromise — it's the design🔗 Find Noelle at the AI Leadership Institute: https://noelle.ai/🔗 Connections:Connect with Noelle: https://www.linkedin.com/in/noelleai/ Connect with Leena Agarwal: https://www.linkedin.com/in/leenaagarwal/#AgenticAI #AILeadership #ResponsibleAI #AIStrategy #NoelleRussell #LeenaAgarwal #ArtificialIntelligence #AIExecution #TechLeadership #MachineLearning