Google

Leading Chrome's global engineering teams focused on activating the best of quality web applications through developers at scale. Our teams work on reducing the friction to test, debug and optimize web applications in the developer's preferred web technology and tooling stack. Portfolio includes products reaching 40M developers serving a Chrome audience of 3B+ users.

My AI specific projects have included: Contributing to Gemini model improvements for coding, Gemini (AI Assistance) in Chrome DevTools across 5 different product launches and Gemini rollouts across multiple developer surfaces in Google Labs.

Head of Chrome Developer Experience

Engineering manager leading organisations working on making Chrome & the web fast. My work involves leading some of our projects to help how quickly Chrome loads pages, shaping APIs, and defining best practices to help large sites deliver great user experiences.

Head of Chrome Developer Insights

Technical Lead and manager for a team in Chrome focused on improving speed and user-experience on the web. Our teams focus on a combination of standards, tooling and library projects for reducing the friction for building success on the web.

Head of Web Developer Ecosystem

O'Reilly

Author of multiple #1 best-sellers including “Leading Effective Engineering Teams” and “JavaScript Design Patterns”

Author

AddyOsmani.com

Between my Medium channel and personal site, the canonical technical articles I author have been viewed more than 17M times in the last few years. I focus on taking complex topics related to Web Performance or the Web Platform and try to distill them down to the core piece of insight that might help others.

Owner and Author

jQuery Project

jQuery team (Bug Triage Team, API Docs Team, Core)

Senior Software Engineer (JavaScript)

Shortsale Engineering

Senior Software Engineer (Web Development, JavaScript)

Empora Group GmbH

Senior Web Developer, primarily responsible for the development of web applications using JavaScript, ActionScript/Flash for widgets and PHP + SSR frameworks for the backend.

Senior Web Developer (JavaScript)

Consultancy and in-house lead roles

This section will detail my previous experience once I get some more free time to update. Covers consultancy across large-scale developers projects requiring JavaScript, PHP, .NET and other front and back-end technologies. For complete details, please request a copy of my resume.

Senior Software Engineer

Addy Osmani is an engineering leader working on Google Chrome. He is the Head of Developer Experience focused on making the web delightful and low-friction for developers to build on. Follow for motivation, productivity and engineering.

Addy Osmani

"21 Lessons from 14 Years at Google"

You can read the full list of 21 lessons here: https://lnkd.in/gvKkjYBi ✍ 

When I joined Google ~14 years ago, I thought the job was about writing great code. I was partly right. But the longer I've stayed, the more I've realized that the engineers who thrive aren’t necessarily the best programmers - they're the ones who’ve figured out how to navigate everything around the code: the people, the alignment, and the ambiguity.

Here are the first 7 lessons from the article:

1. The best engineers are obsessed with solving user problems. It's seductive to fall in love with a technology and go looking for places to apply it. But the engineers who create the most value work backwards: they obsess over understanding the user problem and let solutions emerge from that.

2. Being right is cheap. Getting to right together is the real work. You can win every technical argument and lose the project. The skill isn't being right; it's entering discussions to align on the problem and creating space for others. Strong opinions, weakly held.

3. Bias towards action. Ship. You can edit a bad page, but you can't edit a blank one. The perfect solution rarely emerges from thought alone - it emerges from contact with reality. Momentum creates clarity. Analysis paralysis creates nothing.

4. Clarity is seniority. Cleverness is overhead. Software engineering is what happens when you add time and other programmers. Your code is a strategy memo to strangers who will maintain it at 2am during an outage. Optimize for their comprehension, not your elegance.

5. Novelty is a loan you repay in outages and hiring. Treat your technology choices like you have a small "innovation token" budget. Spend one each time you adopt something non-standard. You can’t afford many. Boring technology has known failure modes; novelty has unknown ones.

6. Your code doesn’t advocate for you. People do. If no one can articulate your impact when you're not in the room, your impact is effectively optional. This isn’t about self-promotion; it's about making the value chain legible to everyone.

7. The best code is the code you never had to write. We celebrate creation, but deletion often improves a system more than addition. Every line you don't write is a line you never have to debug. Before you build, exhaust the question: "What would happen if we just… didn't?"

Read the remaining 14 lessons in the full post :)

#programming #softwareengineering #lifeatgoogle

Director, Google Cloud AI. Best-selling Author. Speaker. AI, DX, UX. I want to see you win.

"Gemini 3 can generate 3D scenes with three.js where you can literally move particles with your hands"

This is a fun prompt you can try in AI Studio, Vertex AI Studio or your favorite vibe-coding tool. 

Prompt:

"Create a real-time interactive 3D particle system with Three.js.

Requirements:

1. Control the scaling and expansion of the particle group by detecting the tension and closing of both hands through the camera.

2. Provide panels that can choose hearts/flowers/saturn/Buddha statues/fireworks and other templates

3. Support the colour selector to adjust the particle colour

4. Particles need to respond to gesture changes in real time. The interface is simple and modern.

5. The interface is simple and modern"

Some optional additions you can play with include "make the object rotate when I swipe" or "make this look even better"

h/t EHuanglu on Twitter

#ai #programming #softwareengineering

New in Chrome DevTools! Debug an app's full performance trace with Gemini! 

The DevTools Performance panel just got a major upgrade with a deeper integration of Gemini! Now you can analyze performance issues faster and more holistically than ever before.

After recording a trace, you can now chat with Gemini about the entire trace, related Performance insights, and even connected field data - all without needing to select specific context beforehand!

Get a full-picture analysis of your page's performance and identify potential bottlenecks: Let Gemini highlight areas of concern before you manually dive into the details.

Once Gemini helps you spot a potential problem, the workflow is smooth:

1. Refine your focus: Easily select a more specific context item - like a single trace event, a specific Flame Chart block, or a Performance insight.

2. Continue the same chat: Keep the conversation going! Gemini will adjust its advice and analysis based on the new, narrow focus, helping you get to the root cause faster.

The power of Gemini is now available for all insights in the Performance > Insights tab. If you see an insight, you can instantly ask Gemini about it for deeper context, explanations, and potential fixes.

This new workflow is designed to help you move from a broad performance overview to a targeted deep dive without ever breaking your flow. Give it a try on your next performance recording!

#ai #programming #softwareengineering

Engineering Leader, Google Chrome. Best-selling Author. Speaker. AI, DX, UX. I want to see you win.

Excited to share my new YouTube series on Software Engineering Effectiveness!

Watch it for free now: https://lnkd.in/gDJVHVgU 

"What separates a good engineer from a truly great one? The secret lies in one powerful word: effectiveness"

I had a fantastic time working with Google for Developers this summer on a brand new show all about Software Engineering Effectiveness at all levels. The first episode is out now and we had so much fun working on it. Subscribe if you would like to be notified of new episodes! :)

It takes a village to pull of such well produced shows like this and I'd like you join me in giving a big shout out to Marcela Villegas Castañón, Caio Avelar, Ella Dobson, Rafael Guetta, M.E. Francis and many others that made this possible!

We hope you find the new series helpful!

#programming #softwareengineering #lifeatgoogle

"How to become an effective senior software engineer" - The T-Shape

My latest free video: https://lnkd.in/gVAAkMzp

As engineers, we often obsess over learning everything. But the most impactful senior engineers I know follow a specific model: The T-Shaped Engineer.

Here is the breakdown:

⬇️ The Vertical Bar (Depth): This is your superpower. It’s that one specific area where you are the undeniable expert. Maybe it’s the billing system, the intricacies of the rendering pipeline or infrastructure. When the team is stuck on this, they look at you.

➡️ The Horizontal Bar (Breadth): This is your context. You don't need to be an expert in everything, but you have working knowledge across the stack. You understand how the database, design system, and CI/CD pipeline connect to your work.

Why does this matter?

Junior engineers often try to widen the horizontal bar too fast without ever digging deep. Senior excellence requires that vertical anchor - that deep expertise that creates trust and reliability.

I hope you're enjoying the "Effective Software Engineer" series so far! It's been great working with Google for Developers on it!

#programming #softwareengineering #lifeatgoogle

Announcing the Gemini Live API is now Generally Available! Here's my JARVIS demo.

For years, building voice agents meant dealing with the "awkward pause" - stitching together speech-to-text models, LLMs, and text-to-speech into a high-latency pipeline.

We are changing that architecture today with the Gemini Live API now GA on Vertex. Powered by the Gemini 2.5 Flash Native Audio model, developers can now build agents that process raw audio, text, and video in real-time over a single WebSocket connection.

I built a simple JARVIS in just 10 minutes using the API on Vertex. It uses Search grounding, Nano Banana Pro and more as you can see in the video.

We've also released a comprehensive (official) Developer's Guide, including a React template and Vanilla JS starter to get you running in minutes. 

📖 Read the Vertex Developer's Guide: https://lnkd.in/gBTbmu7n

🛠️ Get the Code to our official demos (GitHub): https://lnkd.in/gyCwj55E

🤖 Try the Quick Jarvis demo: https://lnkd.in/gCJ5Y7rq

Why is this interesting?

Because the audio is native, we can process tone, cadence, and context instantly. This enables Affective Dialogue. The model has "emotional intelligence" and can detect if you are frustrated, excited, or curious, and adjust its response to match your vibe.

We've also introduced Proactive Audio. This means "barge-in" is smarter than ever. You can interrupt naturally, and the model stops listening and pivots immediately, just like a human would.

But I especially like Robust Tool Use. We are introducing Grounding with Google Search and Functional Calling within the audio stream. This means your agent can pull real-time world knowledge and execute tasks.

And finally, because this is GA, it is enterprise-grade - high availability, multi-region support, and the stability you need to run mission-critical apps.

The Live API with Gemini 2.5 Flash is ready for your production workloads today. It’s fast, it’s grounded, and it’s emotionally aware. I can't wait to see the next generation of conversational agents you build with this.

#ai #programming #softwareengineering #lifeatgoogle

"A free 50-page guide on AI Agents: design patterns, MCP and A2A, context-engineering, multi-agent systems, RAG and more"

At Google we just released a great guide to AI Agents with a clear taxonomy of agent intelligence. It reframes how we should build autonomous systems.

Check where you stand:

❌ Level 0: The core reasoning system. Pure LLM. No tools. Knows history, but can't tell you the weather today.

✅ Level 1: The connected problem solver. "Retrieval." It can use a search tool or query a database (RAG). This is where 90% of enterprise AI lives today.

🚀 Level 2: The strategic problem solver. "Planning." It breaks a mission into steps. It realizes it needs to find a hotel before it can book a car. It performs Context Engineering to manage its own attention.

🤝 Level 3: Collaborative Multi-Agents. "Organization." A Project Manager agent delegates tasks to a Coder agent and a Reviewer agent. They interact via standardized protocols like MCP and Agent2Agent (A2A).

🧬 Level 4: The self-evolving system. "Adaptation." The agent realizes it lacks a tool to solve a problem, so it writes the code for a new tool, tests it, and adds it to its own toolkit.

If you want to move from Level 1 to Level 4, you need to master the Think-Act-Observe loop and implement real Agent Ops (tracing, evaluation, and security).

The full guide covers the architecture for all 4 levels. Highly recommended.

Direct link: https://lnkd.in/gdM9pG2q

#ai #learning #google #lifeatgoogle #programming

Introduction to Agents

"How modern browsers really work" and why it matters 

Read more: https://lnkd.in/gx2sPf_f ✍ 

Modern browsers are some of the most complex pieces of software most of us use every day, yet they are still a bit of a black box for many developers.

I put together a deep dive on how modern browsers work, walking through what actually happens when you type a URL into the address bar and hit enter.

It covers the journey from navigation and networking, through parsing, DOM and CSSOM construction, layout, painting, and compositing, all the way to pixels on screen. Along the way, it also touches on processes, sandboxing, and how modern architectures keep things fast and secure.

Key takeaways

 • Modern browsers are multi-process systems. Each tab, extension, and service runs in a sandboxed process for stability and security.
 • A navigation triggers a full pipeline: networking, HTML parsing, DOM construction, style calculation, layout, paint, and compositing.
 • Rendering is incremental and optimized. Browsers perform work in stages to stay responsive, prioritizing user input and animations.
 • CSS and JavaScript directly influence pipeline cost. Blocked rendering, heavy layouts, and forced synchronous measures cause jank.
 • Compositing is critical. Layers enable the browser to isolate changes and avoid repainting the entire page.
 • Understanding the event loop and task queues helps developers predict responsiveness and avoid long tasks.
 • Performance wins come from working with the rendering pipeline: reduce layout thrash, avoid unnecessary paints, and ship less JS.
 • Shifting mental models from the browser is magic to the browser is a predictable system makes debugging and optimization far clearer.

My goal with this piece is to demystify the browser pipeline so it is easier to reason about performance, jank, memory usage, and how your code maps onto what the browser is actually doing.

If you build for the web, I hope this helps deepen your mental model and gives you a few ideas for what to measure and optimize next.

#programming #softwareengineering #javascript

Treat AI-generated code as a draft. 

It can write the first version, but never outsource the reading. No human review means no reliable trace from behavior back to intent.

When you stop reviewing AI drafts, you stop knowing why the code works at all.

🚀 Introducing Nano Banana Pro! Google's state of the art image generation model 🍌 🍌🍌 ...and 7 mins of demos!

We're excited to launch Nano Banana Pro, a next-generation image generation and editing model built on Gemini 3 Pro. This is a step-change for professional creative workflows and I've really enjoyed using this model lately.

Here is a quick look at what it can do:

- Advanced controls for design workflows: Masked editing, color grading, lighting, and precise aspect ratios, so you can make targeted, pixel level changes.

- Strong prompt understanding: Handles multi layered instructions and intent, so you spend less time iterating on prompts.

- Element preservation: Keeps characters, products, and brand elements consistent across sequences of images and edits.

- Enhanced visual reasoning: Creates rich infographics, diagrams, and UI layouts grounded in Google’s knowledge and your instructions.

- Cleanup and refinement: Powerful tools for retouching and polishing assets so they are ready for decks, campaigns, and production.

- Resolution upscaling: Supports 1K, 2K, and 4K outputs for high fidelity work; in the Gemini app you can generate up to 2K, and in AI Mode in Search up to 1K.

Powered by Gemini 3 Pro’s state-of-the-art reasoning and world knowledge, Nano Banana Pro can follow real-world, highly nuanced instructions. From complex diagrams and multi-step commands to detailed brand and character guidelines, it translates intricate text prompts into accurate, beautiful visuals.

Where you can use it

Across Google products you now have a choice between the original Nano Banana for fast, fun edits and Nano Banana Pro for the most visually sophisticated work.

- Consumers and students: Rolling out globally in the Gemini app when you select "Create images" with the “Thinking” model. 

- Developers and enterprises: Available via the Gemini API and Google AI Studio in paid preview, and integrated into Google Antigravity so teams can build AI first applications and rich UX mockups. Enterprises can also start scaling creation in Vertex AI, and mobile and web developers can call Nano Banana Pro through Firebase AI Logic client SDKs without standing up their own backend.

- Professionals: Image generation in Google Ads is upgrading to Nano Banana Pro to put more advanced creative controls in advertisers’ hands. It is also coming to Workspace customers in Google Slides and Vids to power richer visual storytelling.

- Creatives: Available to Google AI Ultra subscribers in Flow, our AI filmmaking tool, to give filmmakers, creatives, and marketers even finer control over frames and scenes.

Nano Banana Pro is built for anyone who demands precision, consistency, and advanced reasoning from their generative AI tools. Do try it out!

#ai #programming #lifeatgoogle #softwareengineering

taplio