Staff Research Scientist @GoogleDeepMind. Agentic video generation in Gemini Omni.
San Francisco, CA
Joined March 2012
- Tweets222
- Following318
- Followers2.2K
- Likes826
Pinned Tweet
🚀Meet Gemini Diffusion, our first diffusion-based and super fast language model, just announced at Google I/O!🚀
Very excited to be able to share what I've been working on for the past little while with our amazing small team @GoogleDeepMind.
We’ve developed Gemini Diffusion: our state-of-the-art text diffusion model.
Instead of predicting text directly, it learns to generate outputs by refining noise, step-by-step. This helps it excel at coding and math, where it can iterate over solutions quickly. #GoogleIO
Ivana Balazevic retweeted
We are hiring research scientists and engineers🧑🔬 on the Gemini Omni team! Both in the US (Mountain View / SF) and Europe (London / Zürich).
Apply here: google.com/about/careers/app…
Ivana Balazevic retweeted
We’re hiring! This is a rare opportunity to work with the folks that built Veo, Genie, and Gemini Omni — and advance the frontier of world models.
More details below 👇
Ivana Balazevic retweeted
🚨Please repost!🚨
Project to help UNICEF design AI systems
app.unv.org/opportunities/17…
📢UNICEF Egypt Education is looking for an experienced AI Systems Architect to help design an AI-powered Education Intelligence Platform.
🕒 3 months | 1–2 hours/day | Fully remote | Pro bono
Ivana Balazevic retweeted
We’re dropping Gemini Omni: our first step towards a model that can create anything from anything - starting with video.
It combines Gemini’s intelligence with our generative media systems - representing a leap forward in world understanding, multimodality, and editing 🧵
Ivana Balazevic retweeted
If you’re starting a career in research, please consider researching mental illness. Felix fought immensely hard to recover and tried every treatment available to him. they werent enough. There are many people in similar positions now who need help that more research could bring
💔
Today marks 1 year since Felix died by suicide. He was my person. It’s also exactly ten yrs since we danced in a crappy nightclub and madeout the first time. These anniversaries shouldn’t go together, but they do.
@FelixHill84 was brilliant and not ashamed of his mental illness.
Ivana Balazevic retweeted
i am looking for an intern to work on a project on unlearning undesired or unsafe behaviours from pretrained models. please reach out if you’re a good fit :) i’m at neurips till sunday and would love to chat.
Ivana Balazevic retweeted
Are you interested to join @GoogleDeepMind as a student researcher (PhD students)?
I am hiring for a fun project on reward modeling for world models!
If you are interested on video understanding/generation and reward models/critiquing, get in touch and apply below! 👇
Ivana Balazevic retweeted
In Nashville for my last PhD conference 🥲.
Come join today 10:30-12:30 in Hall D (#391) to talk insights, tips and tricks to modify pretraining for representation reuse - scalably.
🚀Joint work w/ @zeynepakata, @dimadamen, @ibalazevic & @olivierhenaff while at @GoogleDeepMind.
Ivana Balazevic retweeted
Writing a STRONGLY WORDED LETTER with Gemini Diffusion and its Instant Edit tab
To add to the buzz of this week, a bit more of a personal update: I’m moving to San Francisco in a few days!
Grateful for the 3.5 amazing years in London and all the wonderful friends and colleagues I’ve met here. ♥️
I’ll be staying at Google DeepMind and happy to chat about Gemini Diffusion and anything else related to AI or otherwise. Excited for this new chapter and more hikes and sunny days!!
🚀Meet Gemini Diffusion, our first diffusion-based and super fast language model, just announced at Google I/O!🚀
Very excited to be able to share what I've been working on for the past little while with our amazing small team @GoogleDeepMind.
Ivana Balazevic retweeted
In 2022, I worked on text diffusion for a bit and wrote a blog post. Since then, people have regularly asked me about scaling diffusion LLMs.
All the while, I was on the first row watching Brendan assemble a cracked team and make it a reality. Now I can stop being coy about it😁
Excited to share what my team has been working on lately - Gemini diffusion! We bring diffusion to language modeling, yielding more power and blazing speeds!
🚀🚀🚀
Gemini diffusion is especially strong at coding. In this example the model generates at 2000 tokens/sec, including overheads like tokenization, prefill, safety filters etc.
Ivana Balazevic retweeted
Got access to Google diffusion.
HOLY SH!T
909 tokens/s ??????
I made a calendar in 3s? 3 fcuking seconds?
Ivana Balazevic retweeted
As I was saying: it's happening
We’ve developed Gemini Diffusion: our state-of-the-art text diffusion model.
Instead of predicting text directly, it learns to generate outputs by refining noise, step-by-step. This helps it excel at coding and math, where it can iterate over solutions quickly. #GoogleIO
Ivana Balazevic retweeted
(3/4) Context-Aware Multimodal Pretraining
By the amazing @confusezius with @zeynepakata, @dimadamen, @ibalazevic, @olivierhenaff
📖 [Paper]: arxiv.org/abs/2411.15099
Ivana Balazevic retweeted
One thing I notice in SF is the blurred lines between work and play time e.g. people performative coding at bars / parties signaling “hustle”. This is extremely peasant-coded and not conducive to great work imo.
History’s elites and greats kept strict borders and took leisure very seriously. It was viewed as a strict requirement to enable great work. Leonardo painted by day and staged court masquerades by night. Churchill drafted war speeches after champagne lunches and silk‑robed baths. Greek philosophers held week‑long wine‑soaked symposia that forged politics. Steve Jobs took long, device‑free walks in the hills that seeded the iPhone.
Work ferociously, build something magical, then close the lid; feast, dance, and wander. Live at the extremes. If you can’t stay off your laptop, skip the party.