Cloud Girl

Why are so many enterprise AI projects failing to deliver real ROI?

In this episode of the Cloud Girl Podcast, we dive deep into the future of enterprise AI, autonomous AI agents, Forward Deployed Engineering (FDE), and what it really takes to transform AI from a productivity tool into a system that changes the economics of a business. Priyanka Vergadia, Cloud Girl sits down with Lak Valliappa, co-founder of Obin.ai and former Google AI leader and researcher to dive into state of AI in 2026

The conversation explores why simply making employees faster with AI often doesn't increase revenue or reduce costs; instead, the bottleneck can simply move somewhere else.

In this interview, you'll learn:
• Why enterprise AI projects struggle to generate real ROI
• AI copilots vs fully autonomous AI agents
• Why “moving the bottleneck” can destroy AI ROI
• How autonomous agents can automate workflows end-to-end
• A real-world loan review example showing how AI can change business economics
• AI agents in finance, asset management, insurance, and other regulated industries
• How to measure AI agent performance and ROI
• AI safety, auditability, bias, compliance, and error tolerance
• What Forward Deployed Engineering really means
• FDE vs Professional Services
• How platform companies should structure FDE teams
• The 3 key roles needed to build enterprise AI solutions
• What companies should look for when hiring AI engineers
• Customer empathy and technical skills for AI engineers
• Why domain expertise matters in enterprise AI
• How AI platforms can scale through specialized engineering teams
• Why building your own AI harness can become valuable intellectual property
• Why finance is a powerful vertical for AI agents

If you're interested in AI, software development, Forward Deployed Engineering, tech careers, developer relations, product storytelling, career growth, and the future of work, subscribe for more.

🚀 Want to go deeper? Check out my cohort:
Storytelling for Forward Deployed Engineers (FDE), $200 off: https://maven.com/pvergadia/storytelling-for-fde?promoCode=FRIENDS200

AI Bootcamp Cohort: https://share.google/CCAmYn5SYfuC87oLT
Use Code: EARLYBIRLD for a 30% Discount

Or send my team an email at [email protected] to discuss any specifics about your situation.

AI is changing faster than ever, but behind every AI model, agent, and inference request is an enormous infrastructure challenge.

In this episode of the Cloud Girl podcast, hosted by Priyanka Vergadia we dive deep into the future of AI infrastructure, AI clouds, GPUs, energy, data centers, inference, open-source models, and physical AI with Crusoe.

Crusoe started in Bitcoin mining and evolved into an AI-focused “neo cloud” built specifically for AI workloads. We explore why AI requires a fundamentally different infrastructure stack, why energy is becoming one of the biggest constraints, and how the industry is moving from simply thinking about GPUs to thinking about the entire journey from “electrons to tokens.”

We also discuss:

• Why traditional hyperscale cloud infrastructure isn't enough for every AI workload
• Why Nvidia GPU density is transforming data center design
• Why energy may be the ultimate bottleneck for AI growth
• AI inference vs. training and why inference could dominate the future
• Open-weight AI models and why companies are moving away from expensive frontier-model contracts
• How serverless fine-tuning can simplify AI model development
• Why AI agents require more than just GPUs
• The growing importance of CPUs, storage, networking, and data pipelines
• How physical AI and robotics will change cloud infrastructure
• Why AI data centers may move closer to the edge
• The future of modular data centers
• Whether vertical integration across energy, data centers, GPUs, and software creates a real competitive advantage
• What CTOs should look for when choosing an AI cloud
• The future of AI infrastructure, specialized silicon, nuclear energy, and inference

If you're interested in AI infrastructure, cloud computing, GPUs, Nvidia, AI agents, inference, data centers, open-source AI, robotics, or the future of artificial intelligence, this conversation is worth watching

What happens when AI writes 90% of your code?

In this episode of The Cloud Girl Podcast, host Priyanka Vergadia sits down with Rajat Gill, Co-founder and CEO of CodeRabbit, to explore how he built one of the world's leading AI-powered code review platforms.

From the rise of AI coding agents to the future of software engineering, Rajat shares why code reviews, not code generation, are becoming the biggest bottleneck in modern software development. He also discusses building AI products, scaling developer tools, hiring in the AI era, and what the next generation of engineering teams will look like.

In this interview, you'll learn:
Why AI code generation is becoming a commodity
Why code reviews are the next billion-dollar opportunity
How CodeRabbit became one of the fastest-growing AI developer tools
How AI is changing software engineering careers
Why product thinking matters more than coding skills
The future of GitHub, AI agents, and software development
How engineering teams are hiring differently in 2026
Why explainability is the biggest challenge in AI-generated code
The future of developers, PMs, designers, and engineering leaders

Whether you're a software engineer, AI engineer, CTO, startup founder, engineering manager, or developer using GitHub Copilot, Cursor, Claude Code, Codex, or ChatGPT, this conversation is packed with insights about where AI development is heading next.

🚀 Want to go deeper? Check out my cohort, Storytelling for Forward Deployed Engineers (FDE), $200 off: https://maven.com/pvergadia/storytelling-for-fde?promoCode=FRIENDS200

Join Our AI Bootcamp Cohort here: https://share.google/CCAmYn5SYfuC87oLT
Use Code: EARLYBIRLD for a 30% Discount

Or send my team an email at [email protected] to discuss any specifics about your situation.

━━━━━━━━━━━━━━━━━━━━━━━━
⏱ TIMESTAMPS
━━━━━━━━━━━━━━━━━━━━━━━━
00:00 – Introduction
01:20 – What is CodeRabbit? Company overview
01:51 – Rajat's background: from Net Sale (observability) to CodeRabbit
03:51 – How the idea for CodeRabbit came about (post-COVID, GitHub Copilot's arrival)
06:28 – Why code review is now the biggest SDLC bottleneck
08:25 – Where AI beats humans vs. where humans still matter
11:52 – Build-more-and-decide-later workflows; risk in mature vs. prototype codebases
14:48 – Rajat's tool of choice: Codex desktop app
18:35 – Team structure fluidity and shifting bottlenecks
19:29 – How hiring has changed in AI era
22:20 – How do we get customers?
26:00 – How do you sustain your pricing while using expensive AI models?
28:27 – What happens behind the scenes during a CodeRabbit review
31:50 – How are you solving the challenge of enterprise codebases being unique?
35:14 – Why adoption is high and low-friction for developers
38:25 – What's your view on the evolution of the AI coding space and the competition?
43:32 – Why validation/review will remain a centralized quality gate
47:07 – Reflecting on the CodeRabbit journey and fundraising story
48:44 – Product roadmap: expanding into Jira, Linear, Slack, and beyond code review
50:30 – Closing thanks and sign-off

━━━━━━━━━━━━━━━━━━━━━━━━

Brand, sponsorship & business inquiries: [email protected]

AI agents are taking over tasks, managing workflows, and accessing your most sensitive systems — but they have no real identity. And that's a massive security risk no one is talking about.

In this episode, Nancy Wang (CTO at 1Password, Venture Partner at Felicis Ventures, and founder of Advancing Women in Tech) breaks down the identity crisis at the heart of agentic AI — and what it means for every developer, cloud engineer, and enterprise security team.

What you'll learn:

→ Why AI agents have no stable identity — and why long-lived tokens are dangerous

→ How MCP (Model Context Protocol) is reshaping identity and access management (IAM)

→ The difference between agent-to-agent (A2A) and human-to-agent authentication

→ How 1Password is protecting 1.3 billion credentials in an agentic AI world

→ What new identity standards are emerging for AI tools, browsers, and local agents

→ The future of cybersecurity when you have 10 agents running under one identity

Whether you're building with agentic AI, working in cloud security, or exploring prompt injection and MCP security risks — this conversation is essential viewing.

🔗 Topics covered: AI agents | agentic AI | MCP security | model context protocol | IAM | AI security | cybersecurity | prompt injection | 1Password | AI tools | cloud security | identity management

👤 Guest: Nancy Wang — SVP Engineering at 1Password, Venture Partner at Felicis Ventures, Founder of Advancing Women in Tech (AWIT)

📺 Cloud Girl is a YouTube channel by Priyanka Vergadia covering cloud, AI, and careers in tech — subscribe for weekly conversations with leaders building at the frontier.

Former Google engineer and Glean founding team member Deedy Das shares unfiltered insights on AI startup investing, what VCs actually look for, and the $100M deal that got away.

In this episode, Deedy reveals:

✅ Why 100% of cold emails to VCs fail (and what works instead)

✅ The 3 things investors actually evaluate (hint: not your pitch deck)

✅ Which AI trends are overhyped vs. game-changing in 2025

✅ How to avoid the biggest mistakes technical founders make

✅ Why he passed on Lovable (now worth $100M+) and what he learned

✅ The truth about fundraising: traction, team, and domain expertise

✅ Current AI investment opportunities: voice AI, RL environments, and more

From building a $7B company (Glean) to investing at Menlo Ventures and running the Anthropic Fund, Deedy breaks down:

- How to get warm intros to VCs (and why cold emails don't work)

- What makes Swedish engineers build better AI products

- The commoditization of startups and how to stand out

- Red flags VCs can spot in 30 seconds

- Why operating experience matters in venture capital

Perfect for: AI founders, technical entrepreneurs, startup operators, and anyone interested in venture capital insights from someone who's been on both sides.

TIMESTAMPS:

00:00 - Introduction

01:29 - Why leave a $7B startup for VC?

05:40 - Biggest mistakes technical founders make

09:08 - What's hot in AI investing right now

14:08 - Infrastructure vs. application layer investments

19:00 - What VCs actually look for (not your deck!)

25:27 - Overhyped vs. game-changing AI trends

29:08 - Why cold emails have a 0% success rate

35:12 - The Anthropic Fund and AI community

40:07 - Risk, failure, and calculated bets

42:13 - The $100M deal that got away (Lovable story)

CONNECT WITH DEEDY DAS:

Twitter/X: [Add handle]

LinkedIn: [Add profile]

Menlo Ventures: https://menlovc.com

ABOUT THE GUEST:

Deedy Das is a Principal at Menlo Ventures investing in early-stage AI, infrastructure, and SaaS companies. He was on the founding team of Glean (now valued at $7B) and previously worked as an engineer at Google and Facebook. He also manages the Anthropic Fund, supporting AI startups building with Claude.

RESOURCES MENTIONED:

- Glean: Enterprise AI search

- Lovable: AI-powered development platform

- Anthropic: AI safety and research

- Scale AI: Data labeling and annotation

- Thinking Machines: RL-as-a-service

#AIStartups #VenturCapital #TechPodcast #StartupFunding #AIInvesting #Entrepreneurship #TechnicalFounders #MenloVentures #GleanAI #StartupAdvice

---

Subscribe for more insights from founders, investors, and operators building the future of AI and technology.

The conversation covers insights from the tech industry, the introduction to Workado, the concept of agentic orchestration and AI, the business impact of agentic orchestration, role transformation and industry trends, and the future of Workado and industry disruption.

Takeaways

• Tech industry insights from a 25-year veteran
• Understanding agentic orchestration and its business impact

Chapters

• 00:00 Tech Industry Insights
• 05:31 Agentic Orchestration and AI
• 13:20 Business Impact of Agentic Orchestration
• 23:29 Role Transformation and Industry Trends
• 29:13 Future of Workado and Industry Disruption

Is a college degree essential for a top-tier tech career? Google Distinguished Engineer and Kubernetes pioneer Kelsey Hightower shares his unconventional path to the top, which started by quitting college after only two weeks.

In this episode of the Cloud Girl Podcast, Kelsey reveals why he found college unmotivating and instead chose to pursue an A+ IT certification. He details his incredible journey from working fast food jobs to landing a contract role with BellSouth, and eventually zig-zagging his way into open source and senior roles at Google.

This interview is packed with invaluable career advice for software engineers, developers, and anyone in tech who feels the traditional path isn't for them. Learn why Kelsey believes his non-traditional path is more relatable than the Ivy League route and how you can define your own metrics for success.

Topics covered:

• The exact moment Kelsey decided to drop out of college.
• How to build a successful career in tech without a degree.
• The critical difference between a "manager" and a "boss".
• Kelsey's surprising perspective on the current AI hype cycle.
• Why community is one of the most powerful tools for career growth.

#KelseyHightower #TechCareer #NoDegree #SoftwareEngineer #Kubernetes #CareerAdvice #Google #Developer