Teaching your AI new tricks.
Santa Clara, CA
Joined June 2016
- Tweets14.1K
- Following906
- Followers362K
- Likes10.8K
Meet Déjà View, a model that turns photos taken from different angles into 3D geometry and camera positions.
By looping a single transformer block, this 117M-parameter model matches or outperforms models up to 10x larger across five benchmarks, using comparable or less compute. Researchers can dial the number of loops to fit their compute budget.
#NeurIPS2026 oral paper + code: research.nvidia.com/labs/dvl…
OpenShell is joining @CloudNativeFdn 💚
Help shape how AI agents access data, use credentials, and execute code. If you’re building agents, we’d love your input!
Check out the project and contribute 👇
Securing AI agents starts at the runtime layer, not just in prompts.
OpenShell is now in the CNCF Sandbox, bringing policy isolation, network controls, and credential management to autonomous agents.
Read more: cncf.io/announcements/2026/1…
#OpenSource #AI
Ask the Experts: Dory Looped Transformer Model | Nemotron Labs xitter.twitterapi.xyz/i/broadcasts/2BlPrVEdk…
This is amazing for Japan! 🇯🇵
Two new NVIDIA Nemotron synthetic datasets for Japanese personas and personal information detection.
Available now on @huggingface:
huggingface.co/datasets/nvid…
huggingface.co/datasets/nvid…
NVIDIA Agentic AI Day にて、合成データセット Nemotron-Personas-Japan V2、および Nemotron PII Japan を公開しました。
■ Nemotron-Personas-Japan V2: 約 1,000 万件の日本人合成ペルソナに、収入や資産などの金融属性を追加。実在の個人情報を使わずに日本の多様なユーザー像を再現できます。
■ Nemotron PII Japan: 医療、金融、行政など 8 分野、約 43 種類の日本語文書を想定した 10 万件規模の完全合成データセット。
PII の位置と種類のアノテーション付きで、検出、マスキング モデルの学習、評価にご活用ください。
👉 nvda.ws/4sTbE1M
Korea’s next wave of AI is taking shape with:
✅ Secure enterprise agents
✅ Expanded model access
@Hyundai Motor Group is exploring NVIDIA NemoClaw and OpenShell, while @AIMSafety_ is integrating OpenShell to help businesses use AI agents securely.
@LG_AI_Research is opening API access to K-EXAONE-2.5, trained on NVIDIA Nemotron datasets.
Explore what’s next for AI in Korea ⬇️
blogs.nvidia.com/blog/korea-…
We’re bringing Korea’s leading open models into production faster! 🏃♂️
New NVIDIA Dynamo deployment recipes support:
✔️ @SKtelecom’s A.X K2
✔️ @LG_AI_Research’s K-EXAONE 2.0
✔️ @UpstageAI’s Solar Open 2
✔️ @Motif_Tech’s Motif-3
These recipes deliver up to 2x inference performance on NVIDIA B200 GPUs.
NVIDIA AI Day Seoul is underway! 🇰🇷
2,000+ developers, startups, and researchers are coming together to test AI agents, explore robotics, and share how they’re putting Korea’s open models to work.
Catch the news, demos, and highlights from Seoul: blogs.nvidia.com/ai-day-seou…
Turn a single image into a 3D world you can explore with Lyra 2.0.
By combining video generation with 3D reconstruction, our researchers create environments along a camera path you choose. Lyra draws on earlier views and learns to correct accumulating visual errors, helping the world stay consistent as you explore and return to familiar places.
Paper, code, and demos: nv-tlabs.github.io/Project-L…
Explore more @NVIDIA research coming to #SIGGRAPHAsia2026, including Lyra 2.0 👀
Watch the reel for more highlights, and find our technical sessions and hands-on labs here: nvda.ws/4sGMWkV
Box + @nvidia:
• Box Context Retrieval, powered by @nvidia Nemotron open models, enables permission-aware retrieval.
• Open-weight Nemotron models in Box AI.
• Collaboration on secure agent execution with Box Mount and OpenShell.
Read more: blog.box.com/box-partners-nv…
Dev Community Live: NVIDIA Developer Live Agent Gauntlet xitter.twitterapi.xyz/i/broadcasts/2pekkBQbk…
Train, Simulate, and Serve World Models with NVIDIA Cosmos Framework and NIM | Cosmos Labs xitter.twitterapi.xyz/i/broadcasts/2ybAVelYO…
Read the full technical deep dive here: nvda.ws/4zuWhz6
Run Polars on a GPU with AWS | Live with cuDF Polars xitter.twitterapi.xyz/i/broadcasts/2dKLYlvjN…
Great to see Nemotron 3.5 Transcribe trending #1 overall on @huggingface
Thank you to everyone for trying it, sharing it, and building with it. 💚
Meet NVIDIA Nemotron 3.5 Transcribe 🎙️
An open-weight, promptable 30B MoE model with ~4B active parameters that brings transcription in 40 languages, speaker attribution, translation, contextual prompting and audio understanding into one checkpoint.
Watch it keep track of who said what when everyone’s talking at once.
1/ 🧵
Accelerating Structural Simulation in Abaqus with NVIDIA CUDA-X xitter.twitterapi.xyz/i/broadcasts/2rJaLZYaM…
We expanded the NVIDIA Agent Toolkit to help developers build agents that use company data to answer questions and make predictions, with controls over what agents can access and do.
Watch how @LilaSciences , @Siemens and @Dassault3DS are using AI agents to design potential therapies, plan factory production and refine vehicle designs.
See what’s new → nvda.ws/4BJ7VrE
Inside Hermes Agent Traces with NeMo Relay | Nemotron Labs xitter.twitterapi.xyz/i/broadcasts/2dKYPmNjm…
AI agents that answer residents’ questions. Vision agents that help reduce traffic delays. Digital twins that simulate 15 years of urban forest growth in 10 minutes.
Our partners are building applications for cities with NVIDIA Nemotron, Cosmos and Omniverse, sharing computing resources across departments and municipalities.
Read about the projects → nvda.ws/3V0KXvx
Have GPUs in Kubernetes? NVIDIA AI Cluster Runtime brings together NVIDIA Dynamo and the infrastructure it needs, with tested component versions and checks to verify your setup.
Follow the walkthrough to configure, deploy and validate your runtime, and catch us at KubeCon North America next week → nvda.ws/4sAxU0a