Addy Osmani

Who they are

Addy Osmani is Director at Google Cloud AI — spent years leading Google Chrome engineering before building agent-skills, an open-source library of production-grade skills for AI coding agents.

Person

Addy came up as an engineering leader inside Google Chrome, where he spent well over a decade shipping web performance tooling and developer experience work before moving to Google Cloud AI around 2025. He marked the transition publicly with a post titled 'Hello, Gemini.' — a rare moment of personal declaration from someone who usually writes about craft. His output is prolific and specific: a book, 'Leading Effective Engineering Teams,' with a companion site at leet.addy.ie; a Substack at addyo.substack.com on engineering management; a dedicated page on agent engineering at addyosmani.com/agents; and the open-source agent-skills repo on GitHub, which he describes as production-grade engineering skills for AI coding agents and which earned an O'Reilly Radar feature. He's delivered 175+ talks globally, including at LeadDev San Francisco on building effective engineering teams, and contributes regularly to LeadDev and The Pragmatic Engineer's newsletter. His writing spans AI agent orchestration, developer productivity, high-leverage engineering practices, and — unusually for this crowd — Stoicism, the subject of his book 'Stoic Mind.' The through-line is long-tenure operator who invests in public craft: every move builds the same compounding library of frameworks for engineering teams.

Company

The freshest move at Google Cloud is the April 2026 announcement of a $750 million partner fund to accelerate agentic AI development — described as the largest single partner investment from any hyperscaler, with Accenture, Deloitte, KPMG, PwC, and NTT DATA named as participants. That same month, Google Cloud Next 2026 framed the moment as the end of the AI pilot era, launching the Gemini Enterprise Agent Platform and the eighth generation of TPUs. In March 2026, Google completed its $32 billion acquisition of Wiz, folding its security platform into Cloud. Sundar Pichai announced planned Alphabet capital expenditure of $175 billion to $185 billion for 2026, with over half of machine learning compute investment directed to Google Cloud. Q1 2026 revenue came in at $20 billion, a 63% year-over-year increase, with a $462 billion backlog.

Market

Google Cloud holds approximately 12-13% of global cloud infrastructure market share in 2026, behind AWS at 28-31% and Microsoft Azure at 24-25%, though it grew faster than both in Q1 2026. Its primary competitors in AI and cloud services are AWS, Microsoft Azure, OpenAI, Anthropic, and Meta. Geopolitical pressure — particularly from EU regulatory frameworks including the EU Cloud and AI Development Act — is creating complex compliance demands, while AI-driven cyberattack risk is pushing enterprise buyers toward integrated security offerings, a gap the Wiz acquisition was designed to close.

How they likely show up

Long tenure inside Google (Chrome through Cloud AI, well over a decade) → thinks in multi-year compounding arcs, not sprint-to-sprint; likely values consistency of direction over frequent pivots.
Published a full book ('Leading Effective Engineering Teams') with a companion site, a Substack, and 175+ public talks → invests heavily in codifying and transmitting frameworks, not just executing them.
Built agent-skills as an open-source side project while operating at Director level → high agency, ships externally to test and validate ideas rather than waiting for internal consensus.
Writes on Stoicism and published 'Stoic Mind' alongside technical content → brings a reflective, long-game orientation to work; probably less reactive to short-term noise than peers.
Role-type is operator, not founder — consistent upward progression within a single organization → optimizes within constraints, plays the inside game, and likely navigates large-org dynamics with fluency.
Publicly marks career transitions with named blog posts (e.g. 'Hello, Gemini.') → comfortable being visible about direction; treats personal brand as a professional asset.

Conversation tips

→ Reference a specific post or chapter — 'I read the high-leverage activities essay on Medium' lands far better than 'I follow your work.' He writes to be read, and he'll know whether you have.
→ Ask about the Chrome-to-Cloud-AI transition on its merits — it's a meaningful shift from developer tooling to AI infrastructure, and the 'Hello, Gemini.' post suggests he's thought carefully about why he made it.
→ If you want to discuss AI agents, come with a specific angle — he has a dedicated page and an open-source project on agent engineering. Generic 'AI is changing everything' framing will fall flat.
→ Stoicism is a genuine intellectual interest, not a brand affectation — if you've read any primary Stoic texts, it's a real opening.

Toolbox

Openers

Open on agent-skills — he built an open-source library of production-grade engineering skills for AI coding agents and got an O'Reilly Radar feature for it. It's the sharpest signal of where he thinks agentic AI actually breaks down in practice.
Reference the '14 More Lessons from 14 years at Google' post — it's a rare personal stocktake from someone who usually writes frameworks, not autobiography, and it anchors the conversation in what he's actually learned inside a large org.
Mention Google Cloud Next 2026's 'end of the AI pilot era' framing — he's a Director on the AI team that shipped Gemini Enterprise and the new TPU generation, so this is his current operating reality, not background context.

Discovery questions

Your agent-skills repo treats production-grade agent behavior as an engineering discipline — what failure modes in real deployments convinced you that framing was right?
After 14+ years inside Google, moving from Chrome to Cloud AI is a significant reorientation — how different is the craft of shipping for enterprise AI versus developer tooling?
Google Cloud Next 2026 declared the end of the AI pilot era — from where you sit on the AI team, what does 'scaling' actually require that piloting didn't surface?

Avoid

Don't open with generic praise about his public profile or follower count — he engages on specific ideas and frameworks, and surface-level flattery will signal you haven't done the reading.

Make it yours