Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile
Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ

@tomaszbednarz

Director of Strategic Researcher Engagement at NVIDIA, SIGGRAPH Asia 2019 Chair, SACAG Chair @siggraphasia, #Simulation #AI #Innovation #Strategy

ID: 19054387

linkhttp://tomasz.matters.today calendar_today16-01-2009 04:16:46

5,5K Tweet

2,2K Followers

4,4K Following

CEDA (@ceda_news) 's Twitter Profile Photo

Introducing 2025 #AILeadershipSummit speaker: Tomasz Bednarz. Tomasz is the Director of Strategic Researcher Engagement at NVIDIA. This October, Tomasz will take to CEDA's stage to discuss the state of #AI both locally and internationally: ceda.com.au/eventsandprogr… ⬅️ 🌐

Introducing 2025 #AILeadershipSummit speaker: Tomasz Bednarz.

Tomasz is the Director of Strategic Researcher Engagement at NVIDIA. This October, Tomasz will take to CEDA's stage to discuss the state of #AI both locally and internationally: ceda.com.au/eventsandprogr… ⬅️ 🌐
Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

⚑️ Gen AI super-resolution is accelerating weather prediction. 50x speedup with NVIDIA Earth-2 & CorrDiff for low-compute, scalable AI models. Faster training & real-time predictions are here. Learn more. πŸ‘‰ bit.ly/47XSfme

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Join our Real-Time Digital Twin (RTDT) Deep Dive livestream! The session will show how to build RTDTs with Omniverse APIs, kit-CAE extension, and NVIDIA PhysicsNeMo, including live code walkthroughs, CAE data import guides (CGNS, VTK, EnSight, NumPy) bit.ly/4rjdsjz

NVIDIA AI Developer (@nvidiaaidev) 's Twitter Profile Photo

Today’s Cosmos Cookbook Special 🍽️ Voxel51's recipe that uses Cosmos Transfer 2.5 to address scarcity in biodiverse data by taking a few existing moth images and generating rich, field-like agricultural scenes. πŸ“– Transform lab-style BioTrove moth shots into photorealistic

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

NVIDIA Omniverse Fixer applies diffusion-based generative techniques to suppress rendering artifacts and recover structural detail in degraded regions of a scene. Learn how to get started bit.ly/3XxjJdN

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Headed to SIGGRAPH Asia? Don't miss learning about how the β€œGenerative Renaissance” is transforming computer graphics & accelerating the rise of physical AI. πŸ‘ bit.ly/44MpRTh

ℏΡsam (@hesamation) 's Twitter Profile Photo

Harvard's just open-sourced their ML Systems textbook. it's extremely practical for not just learning how to build and train models, but to build production systems (the skill that actually matters). topics are cool af: > building autograd, optimizers, attention, and a

Harvard's just open-sourced their ML Systems textbook. it's extremely practical for not just learning how to build and train models, but to build production systems (the skill that actually matters). topics are cool af:
> building autograd, optimizers, attention, and a
Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Build and deploy your own autonomous deep research agent using NVIDIA's AI-Q Research Assistant Blueprint in this new self-paced course. Β  Enroll now: nvda.ws/4iGlWx7 bit.ly/3NnNZFR

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Most OSS models in ’25 use MoE. MoE unlocks massive capability but creates GPU communication bottlenecks. GB200 NVL72 eliminates these with NVLink Switch to link 72 GPUs w/ 130 TB/s bandwidthβ€”delivering 10Γ— better perf/watt vs H200 & 90% lower cost/token. bit.ly/45kFNMO

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Physical AI needs more than dataβ€”it needs world understanding. Announcing open, high accuracy and customizable models like Cosmos Reason 2, Cosmos Predict 2.5, and Cosmos Transfer 2.5 that are pushing the boundaries of embodied intelligence. bit.ly/3LvbRa6

ACM SIGGRAPH (@siggraph) 's Twitter Profile Photo

Become a Student Volunteer Team Leader and guide a team at one of the world’s top conferences for computer graphics and interactive techniques. Coordinate operations, tackle challenges, and help ensure every program runs smoothly. Student Volunteers are the heartbeat of SIGGRAPH

Become a Student Volunteer Team Leader and guide a team at one of the world’s top conferences for computer graphics and interactive techniques. Coordinate operations, tackle challenges, and help ensure every program runs smoothly. Student Volunteers are the heartbeat of SIGGRAPH
Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Last week at #CES2026, we introduced DLSS 4.5. Today, developers can access the newest transformer model through the DLSS Super Resolution Streamline Plugin. Get the full details on DLSS 4.5 + more in our latest technical blog. bit.ly/3NuHmSe

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

NVFP4 is redefining large-scale training. Radical Numerics breaks down NVIDIA’s 4-bit pretraining recipeβ€”covering FP4 math, stability tricks, and Blackwell optimizations for faster, efficient LLM training. radicalnumerics.ai/blog/nvfp4-par… bit.ly/4qFVoPX

Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

As robots take on more dynamic tasks, developers need physically accurate simulation at scale. NVIDIA OSMO enables an end-to-end workflowβ€”from high-fidelity simulation to cloud-scale data pipelines on Azure. bit.ly/4qu0UVI

NVIDIA Europe (@nvidiaeu) 's Twitter Profile Photo

Scientific breakthroughs are moving faster than ever with the support of accelerated computing. Join NVIDIA, University College London, Warsaw University of Technology, University of Liverpool, and King’s College London to explore how accelerated computing and AI are driving

Scientific breakthroughs are moving faster than ever with the support of accelerated computing.

Join NVIDIA, University College London, Warsaw University of Technology, University of Liverpool, and King’s College London to explore how accelerated computing and AI are driving
Tomasz πŸ‡΅πŸ‡±πŸ‡ͺπŸ‡ΊπŸ‡―πŸ‡΅πŸ‡¦πŸ‡ΊπŸ‡ΊπŸ‡Έ (@tomaszbednarz) 's Twitter Profile Photo

Over 1,600 developers. Four weeks and one goal - push physical AI forward. The results from Cosmos Cookoff are in. πŸ“– Full recap β†’ linkedin.com/pulse/cosmos-c… bit.ly/4mfJAlK