I'm an engineer atxAI focusing on multimodal, video generation and world models. My ultimate goal is to build multimodal AGI[0],[1],[2]

• Grok Imagine v1.0
• Grok Imagine v0.9

🤗 Open Source Projects:

• Cosmos: state-of-the-art generative world models
• NeMo DFM: large-scale training and inference framework for diffusion models
• Megatron-LM MoE: Scaling up mixture of experts
• NeMo: scalable training framework for LLMs transformers
• LongVILA: Long-Context VLM for long videos (ICLR'25)
• ActGPT: browser-use agent
• Channel Pruning: Accelerating Very Deep Neural Networks (ICCV'17)
• Epipolar Transformers: Accurate multi-camera pose understanding (CVPR'20)
• AMC: AutoML for model compression (ECCV'18)
• KL Loss: Accurate Object Detection (CVPR'19)
• FSAF: single-shot object detection (CVPR'19)

🎙️ Invited Talks

• NVIDIA Cosmos: World Foundation Model Platform for Physical AI
• Upcycling LLMs into MoE
• Insights on Sora

🤓Grok Heavy Tungsten Cube

PinnedLoading

1. NVIDIA-NeMo/NeMoNVIDIA-NeMo/NeMoPublic

   A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

   Python 16.8k 3.3k

2. NVIDIA/Megatron-LMNVIDIA/Megatron-LMPublic

   Ongoing research training transformer models at scale

   Python 15.2k 3.6k

3. NVIDIA-NeMo/DFMNVIDIA-NeMo/DFMPublic

   State-of-the-art framework for fast, large-scale training and inference of diffusion models

   Python 30 3

4. NVIDIA/Cosmos-TokenizerNVIDIA/Cosmos-TokenizerPublic archive

   A suite of image and video neural tokenizers

   Jupyter Notebook 1.7k 86

5. channel-pruningchannel-pruningPublic

   Channel Pruning for Accelerating Very Deep Neural Networks (ICCV'17)

   Python 1.1k 308

6. KL-LossKL-LossPublic

   Bounding Box Regression with Uncertainty for Accurate Object Detection (CVPR'19)

   Python 721 105