Tulsee Doshi
@tulseedoshi
Followers
12K
Following
1K
Media
31
Statuses
290
Sr. Director & Product Lead, Gemini Models @GoogleDeepmind. @Stanford CS. Opinions my own.
Joined June 2012
Have you tried Gemini 3 Flash yet? What do you think? ⚡️
Frontier-level performance, 3x faster than 2.5 Pro. Here's what to know about Gemini 3 Flash, from @TulseeDoshi ⚡
24
6
138
With Full Self-Driving Supervised engaged, you're 7x less likely to be involved in an accident
0
79
918
Gemini real-time Speech to Speech translation is now in Google Translate on Android! Just use any headphones to hear live translations in 70 languages 🤩 Now @rpandey1234 can translate any Gujarati to English realtime — this will make our next India trip fun :)
24
34
378
Tackle your toughest challenges with Gemini 3 Deep Think mode, live today for Ultra users! 🚀 Gemini 3 Deep Think takes Gemini 3 Pro's strengths even further (and you can see it in the numbers too, with a 🤯 45.1% on ARC-AGI-2). Built for the problems where you need deeper
7
9
148
Help us ripen our roadmap! 🍌
👋🏽 I'm a PM on Nano Banana, let's just call me the Naina that goes banana over images. If you have any feature requests you find a-peel-ing or any big ideas you want to plantain into the team's brains, comment away! cc @nbrichtova @heyitsbeaaaaaaa @oliver_wang2 @m__dehghani
12
1
140
Kids Ride the Ferry Free thru Jan 4, to Midtown & Downtown, NYC. Free Shuttles to Attractions. Watch Video. Click for Routes & Schedules.
0
10
36
As a PM, taking some advice on how to look smart in meetings 😅 Nano Banana Pro is incredible at text rendering + leverages Gemini 3’s reasoning capabilities in the visuals. You can do even more with this model: multiple languages, multiple inputs, and real world reasoning!
21
15
159
Senior Director of Product Management for Gemini @tulseedoshi breaks down the latest on Gemini 3 and Nano Banana Pro ⬇️
75
146
969
Nano Banana Pro needs no introduction, it introduces itself
9
22
159
Nano Banana Pro is here!! 🍌🍌🍌 It builds on all of the best of 🍌 *and* Gemini 3, bringing Gemini 3's reasoning power to leverage real-world knowledge, generate high fidelity text across languages, and build truly visual stunning images https://t.co/nNWtxebT4c
blog.google
Nano Banana Pro is our new image generation and editing model from Google DeepMind.
7
10
117
Properly getting back from my X hiatus, because there's just so much to say about Gemini 3! Got to do one my favorite things and chat with @joshwoodward and @OfficialLoganK about the behind-the-scenes of building our 🔥🔥🔥 model
13
14
170
Gemini 3 created this playable maze in just three prompts 🤯🤯🤯 First, it created a top down Gemini maze, and then we asked it to build an app that allows me to upload a pixel maze, and turn it into a playable Three JS scene. Vibing coding my way through mazes from now on
31
62
764
We dropped a game-changer for on-device AI: EmbeddingGemma. It's the leading model < 500M params & supports 100+ langs; great for private, high-quality RAG across your data—right on your phone, no internet needed. Another week of relentless shipping from Gemini land 🚀 🍌🔒📱
EmbeddingGemma is our new best-in-class open embedding model designed for on-device AI. 📱 At just 308M parameters, it delivers state-of-the-art performance while being small and efficient enough to run anywhere - even without an internet connection.
33
100
1K
The people behind 🍌 magic!
A conversation with some of the research folks behind nano-banana 🍌 (aka Gemini 2.5 Flash Image) on how we got here, what it took to build this model, and where we go next! So much fun to hang with: @19kaushiks @robertriachi @m__dehghani @nbrichtova
2
4
54
@GeminiApp What's particularly amazing? I took just my twitter profile picture & prompted the model to create a trading card of me in a Bollywood film. The fabulous lehenga design and the character consistency? All Gemini. Try it at
gemini.google.com
Meet Gemini, Google’s AI assistant. Get help with writing, planning, brainstorming, and more. Experience the power of generative AI.
3
1
29
🍌 Gemini 2.5 Flash Image, the #1 image editing model, is live in @GeminiApp, AI Studio, & API, with a record-breaking +170 ELO lead! I am ready for my Bollywood audition 💃
Image generation with Gemini just got a bananas upgrade and is the new state-of-the-art image generation and editing model. 🤯 From photorealistic masterpieces to mind-bending fantasy worlds, you can now natively produce, edit and refine visuals with new levels of reasoning,
29
13
662
Suddenly, college life just changed… Here's how to get the FREE PRO VERSIONS of @GeminiApp, @NotebookLM, and more if you’re a university student in the US, Japan, Korea, Indonesia, or Brazil ⬇️
57
108
727
And, everyone's favorite vibe check, creating a pelican with the real beak! https://t.co/QwLSUGNTR2
My notes on Google Deep Think - I don't have a $250/month Ultra account but nickandbro on Hacker News got it to draw a pelican riding a bicycle and the bird actually is recognizable as a pelican!
2
1
15
Creating intricate and complex web apps https://t.co/DNkT2CEabl
I prompted Gemini 2.5 Deep Think: "create a missile command game that incorporates relativity in realistic ways but is still playable." I then asked it to improve the design a few times. The full design & all code & calculations came from the AI, each version ran without errors.
2
2
16
Providing precise instructions for planning agentic tasks https://t.co/PFVSEVd62f
🔥 Gemini Deep Think instructs @AmpCode 🎯 >> you SHOULD use the most intelligent frontier models to generate agent instructions give Deep Think a goal, it creates precise instructions MUCH better at giving precise instructions than I am details on how in 🧵...
1
1
21