🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Hiring (again!) We're looking for a developer advocate specializing in JAX and Keras. This is a really great role if you're interested in going deep in machine learning, and making it easier for developers to train state of the art models on GPUs & TPUs. This role is based in

It'll be great! We're covering the latest in Gemini, Gemma, AI frameworks (JAX, Keras), AI edge & more. Excited to see you all there!

This is great! Fully open foundation model development (including sharing datasets & experiments), in addition to the model weights and code. This makes it reproducible e2e. Bonus: Trained with JAX & TPUs under the hood via the TPU Research Cloud: sites.research.google/trc/about/

“Which part of the AI stack would you say you’re strongest in?” Google: “yes.” #GoogleIO

We’d love to support meaningful work in healthcare. If you’re interested in using MedGemma and need compute support (for tuning or anything else) via the TPU Research Cloud (or other parts of Google Cloud in general), lmk!

Highlights from the AI stage at #GoogleIO2025 yesterday: - Conversation with Demis Hassabis (with special guest Sergey Brin): youtube.com/live/mri4JFkv6… - Overview of our AI stack (from models to frameworks) by Joana Carrasqueira & Josh Gordon : youtube.com/live/mri4JFkv6…

Now with link: developers.google.com/health-ai-deve…

Come read about all the mistakes I made along the way to beating Llama 3.1 8B on 14/19 benchmarks. We trained from scratch, made plenty of wrong turns, and learned a lot.