LinkedIn is better on the app
Don’t have the app? Get it in the Microsoft Store.
Open the appExplore the latest breakthroughs made possible with AI. From deep learning model training and large-scale inference to enhancing operational efficiencies and customer experience, discover how AI is driving innovation and redefining the way organizations operate across industries.
External link for NVIDIA AI
Congrats to Ineffable Intelligence on today’s launch out of stealth. Super excited on our collaboration with their team. More details to come soon.
Introducing Ineffable Intelligence. Led by David Silver, we're assembling the best engineers and researchers in the world to make first contact with superintelligence. We’ll be solving the hardest problems in AI on the way. Come join us. www.ineffable.ai/
📸 New tutorial just dropped Watch to learn how to fine-tune Llama 3.1 with JAX on NVIDIA GPUs – whether single GPUs or multi-GPU and multi-node configurations. 👉 https://nvda.ws/4cwzJVx
Traditional inference wasn’t built for agentic coding. Agentic tools make hundreds of API calls per coding session, often with recomputed context, creating bottlenecks that drive up cost per token. NVIDIA Dynamo rebuilds the stack for agents with: → KV-aware routing → Agent-aware scheduling → Multi-tier caching → Unified orchestration The result: higher cache hit rates, lower latency, and up to 7× more throughput: https://nvda.ws/4u8bjHM
Awesome to see Nemotron-Personas-Korea hit #1 on Hugging Face. Big shoutout to the team and the dev community pushing open datasets forward 💚
🇰🇷💚 Nemotron-Personas-Korea—the very first Korean persona dataset, built with the largest persona PGM ever created (10^46 possible states)—is now officially the 🏆 #1 trending dataset on all of Hugging Face 🥇 I'm especially proud to see a fully Korean dataset at the top ❤️ Thank you all for the incredible support! Featured on the Hugging Face front page alongside OpenAI, Qwen, Kimi, and DeepSeek, this marks another step forward for open-source efforts. After all, data is the fuel for AI. This is a strong validation of NVIDIA’s vision for Sovereign AI and Data Designer. Stay tuned, more exciting projects are on the way! 🔥🚀 https://lnkd.in/gBVK-HtW
📊 Day 0 performance is here: DeepSeek-V4-Pro running on NVIDIA Blackwell Ultra. Using vLLM's Day 0 recipe, we’ve captured the initial performance Pareto for DeepSeek’s flagship 1M long-context model. This curve highlights the baseline for balancing AI factory throughput with real-time user interactivity. This is just the starting line. Expect these numbers to climb as we optimize the full co-design stack, including: • NVFP4 & Dynamo • Optimized CUDA kernels • Advanced parallelization techniques and beyond Read the full technical deep dive: https://nvda.ws/4vRNhm9
Happy Friday! We just put DeepSeek-V4-Pro up on build.nvidia.com. It’s the world’s largest open source model at 1.6T parameters, and you can run it for free running on NVIDIA Blackwell GPUs. Try the NVIDIA NIM API → https://lnkd.in/ghygMYQ4
✨ DeepSeek-V4 is here — a million-token context, 1.6T parameter powerhouse optimized for agentic workflows. Out of the box, on DeepSeek-V4-Pro, NVIDIA Blackwell Ultra delivers over 150 TPS/user interactivity for agentic workflows. And we’re just getting started. Expect these performance figures to climb higher as we implement Dynamo, NVFP4, and advanced parallelization techniques. Start building today with SGLang and vLLM.
Gemma 4 introduced a powerful new family of native multimodal and multilingual models that scales across the full spectrum of NVIDIA hardware - from Blackwell in the data center to Jetson at the edge. In this stream, we’re going hands-on with the DGX Spark to see how it can amplify Gemma 4’s features, including massive 256K token context window and native vision/audio capabilities. Bring your questions! We’ll have experts from NVIDIA and Google DeepMind.
We just wrapped Nemotron Dev Days Korea — a two-day intensive bringing together 200 elite developers to push the boundaries of sovereign AI. In this session we break down what happened, what we learned, and what it means for builders everywhere. We'll open with a recap of the Korea hackathon: the three tracks (Creative Use-case, Domain & Persona Tuning, and Synthetic Data Generation), standout submissions, and a closer look at the winning teams and what they built. You'll hear directly about how participants used Nemotron models, NeMo-RL, NeMo Curator, and the broader NVIDIA stack to build production-ready pipelines in under 36 hours. Then we zoom out to the bigger picture — sovereign AI model builders across Korea and the region are increasingly turning to Nemotron as the foundation for domain-adapted, culturally-aware models. We'll walk through how teams are leveraging the Nemotron SDK, curated datasets, and NVIDIA's post-training recipes (CPT, SFT, RL) to build models that go well beyond out-of-the-box performance. We close with a live Build-a-Claw segment — a step-by-step walkthrough of setting up NemoClaw to deploy your own autonomous agent, using the same blueprint that hackathon teams had access to. What you'll learn: - Highlights from Nemotron Dev Days Korea — tracks, results, and winning builds - How sovereign AI builders are using the Nemotron SDK, datasets, and post-training recipes in real-world deployments - What "Domain & Persona Tuning" looks like in practice for Korean-language and domain-specific models - How to get started with Build-a-Claw: installing NemoClaw and deploying your first autonomous agent
What’s possible with Gemma 4 + DGX Spark? Tune in Friday as NVIDIA and Google DeepMind experts walk through vision translation, long-context document Q&A, and real-time code gen demos live. Bring your questions. 📆 Friday, April 24 at 11:00 AM PDT Add to schedule ➡️ https://nvda.ws/4vP77yg