Thoughts

A little delayed, but I really wanted to take a moment to share some of the incredible work my team recently shipped. We brought Gemini to TV, launched TCL's new presence-aware TV, and rolled out an updated navigation model.

Even though we didn't start the Gemini-on-TV project, we co-owned this with David Gardner mid-flight and got it across the finish line. Building a chat UI on TV is… not a solved problem. We had to rethink everything: how long LLM responses should be so they feel “just right” on a big screen, and how to optimize voice plate animations to run smoothly on TV hardware (which is way harder than it sounds).

On top of that, the new navigation model makes the experience simpler while creating space for the broader Google ecosystem—Google Home, Gemini, and more.

Massive kudos to Advait Kalakkad, Zhengnan Zhao, Stacy Li, and Alvaro Delgado Perez for bringing these features to life. I'm incredibly proud of what we pulled off with such a tiny team.

Not long ago, I was chatting with a senior engineering manager. During our conversation, I mentioned how I built an extension for VS Code. To my surprise, they had never heard of VS Code before.

In the development world, VS Code is one of the most widely used code editors. But this made me realize: even experienced professionals can sometimes get stuck in their own bubble. In this case, the manager likely hadn't coded in years and was mostly familiar with Android Studio from their Android background.

The takeaway? No matter how experienced we are, it's important to stay curious and aware of what's happening outside our immediate domain. Whether it's new tools, programming languages, or emerging trends, being open to learning keeps us sharp and relevant.

I've noticed this in other areas too—backend engineers hesitant to try frontend work, or designers hesitant to code. But stepping out of our comfort zones is how we grow.

Recently, I started learning Rust. It's been challenging, but also incredibly rewarding to explore something outside my usual space.

Last week, I wrote about having a bias towards action—and a few people reached out asking, “What other qualities do you really value in teammates?”

Here are 6 traits I look up to and try to practice myself:

1. Bias towards action
   They don't wait for the perfect plan—they just start. They move things forward when others hesitate.

2. Attention to detail
   Nothing is too small to fix. Whether it's a misaligned slide or a mismatched color, they care enough to make it right. That mindset compounds.

3. Cautious ambition
   It's easy to be ambitious. Harder to be ambitious and realistic. The best people know how to read the room, aim high—and still stay grounded.

4. Respect for the problem
   They don't rush to solutions or hide behind “first principles.” They pause. Think. Put in the time a real problem deserves.

5. Consistency
   Not once. Not twice. But every single time. These are the people you can count on.

6. Persistence
   Most important of all. They push through the messy middle. They finish.

I'll be honest: I'm not great at all of these either. Few of us are. But I wrote them down to remind myself what I value—and who I aspire to be.

Over the past 9 years at Google, I've had the privilege of working with people from all kinds of backgrounds—engineers, designers, PMs, researchers, TPMs, QA. Some were brilliant problem solvers. Some were incredibly hardworking. And some were just okay.

But there's one quality I consistently admire in the people who really make things happen: a bias towards action.

In many meetings, I've seen important but messy problems come up—things that don't clearly fall into anyone's role or team. Everyone agrees it's important. There's talk like, “we should do this”, or “this is a great idea, let's do it”, or “someone needs to take this on.” But once the meeting ends, nothing moves. Emails get passed around. People wait for someone else to pick it up.

The people I see succeeding aren't waiting. They jump in. They get the ball rolling—even when the work isn't part of their official job description. They don't wait for perfect alignment or a green light from leadership. They just start. And if it doesn't work out? They adjust.

When I first joined the Google TV team, my manager Michael DelGaudio told me something I've carried with me ever since: Sid, It's better to say sorry than ask for permission.

That mindset has helped me move faster, take initiative, and ship things that might not have happened otherwise.

Of course, you still have to use common sense and make sure you're not jeopardizing team goals, stepping over others' work or causing dissonance—but progress often starts with someone choosing to act.

Moral of the story: Just do it. The people who take initiative are the ones who move things forward.

I recently interviewed a candidate for a UX Design role at Google. Their portfolio website was beautifully designed and caught my attention, so I asked them how they built it.

They gave me a detailed answer about how they designed and developed the site from scratch over time using Framer. But as I was reviewing their portfolio, I noticed something—they had forgotten to remove the link to the original template they used. It was clear they had only made minimal changes—maybe 10%—while the rest remained unchanged from the original template.

And that's okay, using templates is totally fine. I would have appreciated a simple, honest answer like: "I found a great template online and customized it to fit my needs." That kind of honesty builds trust. But trying to take full credit for someone else's work? That only raises red flags.

Moral of the story: Be upfront about what you did (and didn't) do. It builds more credibility and respect than overstating your contributions. Authenticity goes a long way.

At some point in your career, there's someone who takes a chance on you. Someone who gives you that first opportunity—the job, the project—that changes everything. Today, I want to take a moment to thank the people who took a chance on me.

After my undergrad, like many engineering students in India, I landed a job at a consulting company that hired in bulk. In fact, they hired 250 people from my college alone! At the time, I wasn't sure what I wanted to do or where my interests really lay.

One day, I cold-emailed a team that I found interesting, and Ranjithkumar Rajarethinam took a chance on me. After a quick 30-minute interview, he offered me a role at the Web 2.0 Innovation Labs. That opportunity, and the amazing colleagues I worked with there, changed my career trajectory. They encouraged me to pursue my master's degree in HCI, and I'm so grateful they did.

While working full-time, I prepared for my GRE and TOEFL—but my scores weren't stellar. Still, I reached out to several professors, and one more time in my life Prof. Gary Hsieh took a chance on me. Despite not so high scores, he recommended my admission to University of Washington's HCDE program. That gave my career the turbo boost it needed.

Once someone gives you an opportunity, it's up to you to make the most of it. After my master's, that path eventually led me to Google.

I never imagined I'd move to another country, let alone work at Google. It felt like a distant dream. But here I am, because people took a chance on me—and I gave it everything I had in return.

So today, I want to publicly thank Ranjithkumar Rajarethinam and Gary Hsieh. I've thought about expressing this with them many times, but I've always fallen short of the words to truly express how grateful I am. This is my small attempt.

I studied computer science in undergrad and never imagined I'd become a UX designer. My first job was on a front-end team, doing a lot of web development. Over time, I became more interested in user experience, so I pursued a master's in HCI.

While applying for jobs after graduation, I had a tough choice to make: UX design or front-end engineering. Listing both on my resume didn't get much traction. So I made a choice and committed to the UX design path. But here's the thing—I thought my software engineering days were over. I was wrong.

At Google, even though I'm a UX designer by title, I've found plenty of ways to keep my engineering side alive. It started with fixing small bugs. Over time, I contributed to production code, built countless prototypes, and now I'm Kotlin readability certified — a recognition given to engineers who consistently write clear, idiomatic, maintainable code that meets Google's standards. Today, I lead our TV UI infrastructure while still designing. The builder in me never left.

Moral of the story: When there's a will, there's a way. If you're passionate about something, go for it—you'll find a way to make it happen.

PS: CL stands for Changelist equivalent to git commit

Today I want to share a quick tip!

Ever needed to download a GIF or an image embedded in a Google Doc — maybe an animated mockup you want to use elsewhere? If you're in the Google ecosystem, you can copy-paste it easily. But if you want to download it, things get tricky.

Here's a simple trick:

1. Go to File → Download → Web Page (.html zipped)
2. Unzip the folder
3. You'll find your GIF inside images folder! 🎬

Hope this helps! Let me know if you've tried it.

At Google TV, we support roughly a 100 languages — some spoken by millions, others by just a few thousand. But no matter how big or small the audience, every language matters.

Google has a robust internal translation pipeline where expert linguists ensure that every string in our products are translated with context and accuracy. However, when it comes to localized images its not that straightforward — there are agencies that provide this as service for marketing materials, but they can be expensive and impractical at scale.

One of our recurring challenges is our quarterly feature update, which includes an image alongside descriptive text. A shortcut would have, stick to one language or remove text from images entirely. But that wouldn't be inclusive. Instead, we aimed for a solution that:

1. Uses real product screenshots in the correct language
2. Ensures translations match the actual UI (no plugin-generated inconsistencies)
3. Scales across 100+ languages without manual effort

Some options we exported, Manually switch languages and take screenshots (time-consuming), automate with scripting (tricky but possible), or find an innovative approach.

So, I created a Figma script using the Scripter plugin (link in comments):

• Pulled the correct translations from our internal system (JSON format)
• Prepared RTL and LTR assets in Figma
• Mapped text components to their corresponding translation IDs
• Ran the script—voilà! 🎉 100 localized images generated in seconds

The initial setup took a few hours, but after that, the creation process became effortless. We could have taken the easy route—but we chose to do the right thing. Accessibility and inclusivety matter.

By no means all the translations are 100% perfect in every language, but we will strive to get it right and fix things as we go. That is why I love working at Google TV as its fun and challenging! :)

So, what makes a designer truly great? It's attention to details. Attention to detail is the single most important quality that sets great designers apart. I believe this focus on detail translates to persistence, and persistence ultimately leads to success.

Optical balancing icon is one such thing, Many designers might overlook such subtle refinements, but at Google TV, we pay attention to these nuances. Recently, I updated all remote icons to be optically centered rather than mathematically centered. By no means Google TV is perfect but we are polishing one thing at a time, eventually we want to build a product that we are proud.

CES wrapped up last week, and I'm thrilled to share some exciting updates from the event! Google TV unveiled the next generation of devices featuring cutting-edge Gemini models, far-field microphones, 4K image support, and proximity sensors, all designed to make your home smarter and more helpful than ever.

I'm incredibly proud that two of these four features showcased at CES were designed by my team! 🎉

The response has been phenomenal, and the glowing reviews this week are a testament to the innovation and hard work behind these devices. Here are a couple of highlights:

💬 “From simply using your presence to turn on your TV to using voice commands to control it, I could see how the new 'smarts' offered a completely hands-free way of interacting with the display…” — TechRadar

💬 “All the features seem to be centered around making your interactions with your TV more seamless and incorporating it into your daily life more than just watching content.” — ZDNET

A huge thank you to the entire team and everyone who made this possible. Here's to building even more seamless and smarter experiences for everyone in 2025 with Gemini and more 🚀

Sometimes, it's fun to build something just because. I spent days—if not months—searching for ways to integrate a spell-checker into the Monaco code editor (the core behind VS Code).

You might wonder, "Don't we already have spell-checkers everywhere?" We do, but Monaco doesn't natively support them due to how it's designed.

For many days I tried server-side solutions like Grammarly and LanguageTool, but they were slow and costly. Then I discovered Harper, an early-stage Rust project. Its speed impressed me, so I forked it, optimized it for my needs my fork, and successfully integrated it into my tool.

The thing is, I had never written a line of Rust before. With AI and just an hour of tutorials, I learned enough Rust to write code, compile it into WASM, and run it directly in the browser—no server needed. It felt incredibly productive.

If you're curious, check-out the tool I've been building for fun: https://stubby.io. It's still in beta, but I'd love your feedback!

Around 22K people use the plugin I built for myself a while ago. The problem you are facing may be one that many others face as well. I'm happy that it has really helped so many people.