Accelerating data science as eng director @NVIDIA RAPIDS. Previously DL for self driving cars, medical imaging, health data, econometrics, and kernel hacking.
Berkeley
Joined May 2011
- Tweets936
- Following228
- Followers334
- Likes785
John Zedlewski retweeted
Our general-purpose coding agent just scored 100% on the ARC-AGI-3 interactive reasoning benchmark.
NVIDIA AVO completed all 183 levels across all 25 public environments, figuring out what to do with no instructions, explicit rules, or stated goals.
John Zedlewski retweeted
Nvidia's GPU UMAP now scales to multi-GPUs, allowing it to run on hundreds of millions of samples in mere minutes.
developer.nvidia.com/blog/ru…
John Zedlewski retweeted
My team open sourced a set of agent skills to run an entire @kaggle competition workflow from plain language. We ship an easy to install plugin for your favorite coding agent.
👉 Try it now: github.com/NVIDIA/nvidia-kag…
John Zedlewski retweeted
Your AI is bottlenecked by your data. Our Dell AI Data Platform: 12x faster vector indexing, 3x faster processing with Lightning FS (the fastest parallel file system in the world), feeding GPUs at 150 GB/s per rack. 🚀
dell.com/en-us/blog/ai-at-sc…
John Zedlewski retweeted
Nvidia just crossed Google as the biggest org on @huggingface with 3,881 team members on the hub.
I'm officially calling it:
Nvidia is the new American king of open-source AI!
John Zedlewski retweeted
Quoting Jensen: "All of these platforms are processing DataFrames. This is the ground truth of business. This is the ground truth of enterprise computing. Now we will have AI use structured data. And we are going to accelerate the living daylights out of it."
Polars DataFrames are at the core of the AI revolution.
youtube.com/watch?v=jw_o0xr8…
John Zedlewski retweeted
Truly awesome to see Jensen in his GTC keynote present @nvidia's joint partnership with @IBM to use cuDF to accelerate open source Velox and the Presto SQL engine in watsonx, showcasing how @Nestle achieves 5X faster queries for 83% less cost.
Amazing collaboration between the Nvidia and IBM teams to make this a reality, and we're just getting started.
John Zedlewski retweeted
NVIDIA won Deep Research Bench with NVIDIA AI Q and Nemotron 3 models. A blog is under way.
Very proud that a cross NVIDIA effort led to top positions in Deep Research Bench I and II. DRB-I evaluates overall report quality: comprehensiveness, insight, instruction-following, and readability whereas DRB-II is more difficult and tests for information recall, analysis, and presentation. This was led by David Austin and @raja_biswas from my team. I contributed a bit. We will share more ASAP.
huggingface.co/spaces/muset-…
agentresearchlab.com/benchma…
John Zedlewski retweeted
Very proud to share that @JiweiLiu, @MJeblick , and @jackyu815 from my team just won DABStep benchmark with an agent that learns from the tasks it solves.
We'll share more details ASAP.
huggingface.co/spaces/adyen/…
John Zedlewski retweeted
I love this view, with us (NVARC, with Ivan Sorokin) being on the right efficiency!
We also verified that GPT-5.2 Pro (High) is SOTA for ARC-AGI-2, scoring 54.2% for $15.72/task
(Due to API timeouts, we were unable to reliably verify GPT 5.2 Pro X-High on ARC-AGI-2)
All verified GPT-5.2 family scores:
arcprize.org/leaderboard
John Zedlewski retweeted
NVIDIA #CUDA-X libraries are accelerating scientific breakthroughs with the power of #AI.
🧬 NVIDIA cuEquivariance takes protein structure analysis from months to minutes, and RAPIDS-singlecell lets researchers process massive volumes of genomics data in seconds.
🔋 NVIDIA ALCHEMI rapidly accelerates battery and OLED discoveries, turning years of lab work into days.
🌎 NVIDIA Earth-2 and NVIDIA PhysicsNeMo enable realistic weather forecasts in seconds, helping scientists predict disasters with AI-driven precision.
Watch here to learn more 👇
or on YouTube ➡️ youtube.com/watch?v=vNhbqYts…
John Zedlewski retweeted
🎊 Congrats to our NVIDIA cuOpt team who just won the 2025 COIN-OR Cup 🏆, awarded for top #opensource achievements in decision optimization and operations research.
cuOpt runs LP, MIP, and VRP workloads on NVIDIA GPUs, delivering fast, scalable results with a robust open codebase.
Get started 🖥️ github.com/NVIDIA/cuopt
John Zedlewski retweeted
HUGE
> we have a new #1 on the
> MTEB Embedding Benchmark Leaderbord
llama-embed-nemotron-8b
> beats Gemini, Qwen3, Linq, GTE
> fits on ONE RTX 5090
> 4096 dims, 32k ctx
> 69.46 avg across tasks
> multilingual bi-encoder, zero-shot across retrieval, rerank, STS, classification
John Zedlewski retweeted
🔥 Join us at #SciPy2025 for our tutorial, "Scaling Clustering for Big Data: Leveraging RAPIDS cuML" with Allison Ding!
🚀 Come learn to implement, optimize, and scale clustering algorithms effectively unlocking faster, deeper insights in your machine learning workflows.
John Zedlewski retweeted
⚡cuML accelerations for ML workflows for scikit-learn, UMAP and HDBSCAN -- with zero code changes required. cuML is up to 60x faster for UMAP and up to 175x for HDBSCAN over traditional CPU-based implementations.
Read the intro to GPU-accelerated UMAP and HDBSCAN workflows: odsc.medium.com/accelerating…
John Zedlewski retweeted
🎉 Huge congrats to our NVIDIA team “NemoSkills” for winning the AIMO-2 Competition 🏆 on @Kaggle.
Their system solved 34 out of 50 problems in just 5 hours using 4 L4 GPUs. 🔢✨⏱️
kaggle.com/competitions/ai-m…
How? A powerhouse squad—Christof Henkel, Darragh Hanley, Ivan Sorokin, Benedikt Schifferer, Igor Gitman, Shubham Toshniwal, and Ivan Moshkov—combined expertise in LLM training, data generation, and KGMON insights.
Their final test set performance even beat earlier public results, proving they don’t overfit.👏
More details on the final solution are coming soon.💡
@kagglingdieter, @gonedarragh, @igtmn, @shubhamtoshniw6, and @i_vainn
John Zedlewski retweeted
🎊 Llama Nemotron Ultra 253B is here 🎊
✅ 4x higher inference throughput over DeepSeek R1 671B
🏆Highest accuracy on reasoning benchmarks:
💎 GPQA-Diamond for advanced scientific reasoning
💎 AIME 2024/25 for complex math
💎 LiveCodeBench for code generation and completion
Try as #NVIDIANIM ➡️ build.nvidia.com/nvidia/llam…
Technical deep dive ➡️ developer.nvidia.com/blog/bu…
John Zedlewski retweeted
🚀 The Colab team collaborated closely with @nvidia to deliver day 1 compatibility for NVIDIA cuML's Zero Code Change ML Acceleration. Now, you can experience significant speedups in your machine learning workflows in Colab with no code modifications! Example notebook below 👇
youtu.be/cIJsVq8CPys?feature…
John Zedlewski retweeted
👀 See how RAPIDS cuML 24.10 makes GPU-accelerated #UMAP even faster and scalable to larger-than-GPU-memory datasets, reducing runtime from hours to minutes on 150GB+ datasets.
➡️ nvda.ws/3NYP2cs
Dive into the technical walkthrough to learn more.
