Nithin Chalapathi

Special thanks to @sumanthrh, @CyrusHakha, and @pcmoritz for making this possible! 🙂

As of this release, the following features are supported: + Multi-modal Tinker recipes (including SFT workloads) + Single image, multi-turn VLM RL with tool calling + Multi-image agentic RL + LoRA and full-fintuning + FSDP Support Try it out now! Docs: docs.skyrl.ai/docs/tutorials… Github: github.com/NovaSky-AI/Sky… Blog: anyscale.com/blog/vision-la…

SkyRL now supports end-to-end vision-language post-training, from SFT to agentic RL, and adds vision model support to SkyRL’s Tinker interface! Existing multimodal cookbooks, e.g. VLM classification, work out of the box:

Encourage everyone to take a look or apply! I've been very fortunate to work with @Ritwik_G who has been an incredible advisor / mentor 😁

where is your equivariant god now? x.com/ask1729/status…

1/ Can molecular AI move past hard-coded Graph Neural Networks and embrace scalable Transformers that discover molecular structure on their own? We demonstrate that you can train a 1B parameter Transformer model without any graph priors or physical inductive biases. And surprisingly, not only can it maintain competitive performance under equal compute on the Open Molecules 2025 dataset… it’s faster than a 6M parameter equivariant GNN, and exhibits scaling laws that don’t saturate. We use this as a starting point to investigate emergent internal representations, and find that it adaptively discovers molecular structure! Check out the interactive demo on our website: tkreiman.github.io/projects/graph… And our paper: arxiv.org/abs/2510.02259 In collaboration with @tobykreiman, @YutongBAI1002, Fadi, Elizabeth, and @EricQuCal. Here’s a video showing how the Transformer learns distance-aware attention patterns (purple gradient) that adapt to atomic environments 👇

@benpearce Same - but I appreciate that I got a little booklet with candidate blurbs

@benpearce no omg this feels like an alien dish, not something out of utah

@benpearce in absolute shock