1. X
  2. Max Zhdanov
Log inSign up
Max Zhdanov
666 posts
Max Zhdanov profile banner
user avatar

Max Zhdanov

@maxxxzdn
intern @GoogleDeepMind, phd @AmlabUva, prev. intern @cusp_ai
London
maxxxzdn.github.io
Joined March 2022
487
Following
2,280
Followers
RepliesRepliesMediaMedia
  • Pinned
    user avatar
    Max Zhdanov
    @maxxxzdn
    May 20
    🌍Today we release Mosaic, a probabilistic weather model that shifts the Pareto frontier of ML weather forecasting. It matches the skill of state-of-the-art models while generating a 24-member, 10-day global forecast in under 12 s on a single H100. Thread!
    Image
    00:00
  • user avatar
    Max Zhdanov
    @maxxxzdn
    Aug 6
    AI that benefits humanity🌍🚀
    user avatar
    Google DeepMind
    @GoogleDeepMind
    Aug 6
    Predicting cyclones accurately can help save lives - and every hour of lead time counts. Published in @Nature, our AI model WeatherNext achieves state-of-the-art accuracy in forecasting a storm’s track and intensity, giving us a critical extra 24 hours to prepare on average. 🧵
    Image
  • user avatar
    Max Zhdanov
    @maxxxzdn
    Jul 17
    I have been using Weather Lab to plan my daily activities for some time now, and it's amazing!
    Image
    Image
    00:20
    user avatar
    Peter Battaglia
    @PeterWBattaglia
    Jul 17
    We're announcing a major update to Weather Lab, our interactive website for sharing Google’s AI weather models from @GoogleDeepMind and @GoogleResearch: deepmind.google.com/science/weathe…
  • user avatar
    Max Zhdanov
    @maxxxzdn
    Jul 8
    I unfortunately am not going to ICML this year, but if you are interested in sparse attention, weather forecasting or high fidelity deep learning, I would love to chat! Attached is Mosaic's learned (sparse) attention pattern whose physical consistency is cool to see.
    Image
    00:00
    Image
    01:21
    user avatar
    Max Zhdanov
    @maxxxzdn
    May 20
    🌍Today we release Mosaic, a probabilistic weather model that shifts the Pareto frontier of ML weather forecasting. It matches the skill of state-of-the-art models while generating a 24-member, 10-day global forecast in under 12 s on a single H100. Thread!
  • user avatar
    Max Zhdanov
    @maxxxzdn
    Jul 3
    the most important video of the year
    user avatar
    Xenova
    @xenovacom
    Jun 13
    I gave Fable 5 one job: write custom WebGPU kernels for Gemma 4 inference. It climbed to 84 tok/s, then hit a wall, insisting further optimization was impossible. Hours later, Anthropic rolled back invisible LLM development safeguards, and it hit 255 tok/s. The next day,
    Image
    00:00

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Advertisement
Advertisement