Serving Large Language Models (LLMs) at scale is a massive engineering challenge because of Key-Value (KV) cache management. As models grow in size and reasoning capability, the KV cache footprint increases and becomes a major bottleneck for throughput and latency. For modern Transformers, this cache can occupy multiple gigabytes. NVIDIA researchers have introduced KVTC (KV…
NVIDIA has released Nemotron-Nano-3-30B-A3B-NVFP4, a production checkpoint that runs a 30B parameter reasoning model in 4 bit NVFP4 format while keeping accuracy close to its BF16 baseline. The model combines a hybrid Mamba2 Transformer Mixture of Experts architecture with a Quantization Aware Distillation (QAD) recipe designed specifically for NVFP4 deployment. Overall, it is an ultra-efficient…
In this tutorial, we explore how federated learning behaves when the traditional centralized aggregation server is removed and replaced with a fully decentralized, peer-to-peer gossip mechanism. We implement both centralized FedAvg and decentralized Gossip Federated Learning from scratch and introduce client-side differential privacy by injecting calibrated noise into local model updates. By running controlled experiments…
Maia 200 is Microsoft’s new in house AI accelerator designed for inference in Azure datacenters. It targets the cost of token generation for large language models and other reasoning workloads by combining narrow precision compute, a dense on chip memory hierarchy and an Ethernet based scale up fabric. Why Microsoft built a dedicated inference chip?…
We implement an advanced, end-to-end Kornia tutorial and demonstrate how modern, differentiable computer vision can be built entirely in PyTorch. We start by constructing GPU-accelerated, synchronized augmentation pipelines for images, masks, and keypoints, then move into differentiable geometry by optimizing a homography directly through gradient descent. We also show how learned feature matching with LoFTR…
Tencent Hunyuan has open sourced HPC-Ops, a production grade operator library for large language model inference architecture devices. HPC-Ops focuses on low level CUDA kernels for core operators such as Attention, Grouped GEMM, and Fused MoE, and exposes them through a compact-C and Python API for integration into existing inference stacks. HPC-Ops runs in large…
DCAI, the AI infrastructure company that operates the Gefion AI Supercomputer, has achieved ISO/IEC 27001 certification, the globally recognized standard for information security management systems (ISMS). This milestone underscores DCAI’s commitment to delivering secure, compliant, and enterprise-grade AI infrastructure. ISO/IEC 27001 is the global benchmark for information security. The certification was awarded by an independent firm after…
Thinking Machines Lab has moved its Tinker training API into general availability and added 3 major capabilities, support for the Kimi K2 Thinking reasoning model, OpenAI compatible sampling, and image input through Qwen3-VL vision language models. For AI engineers, this turns Tinker into a practical way to fine tune frontier models without building distributed training…
Thinking Machines Lab has moved its Tinker training API into general availability and added 3 major capabilities, support for the Kimi K2 Thinking reasoning model, OpenAI compatible sampling, and image input through Qwen3-VL vision language models. For AI engineers, this turns Tinker into a practical way to fine tune frontier models without building distributed training…
OpenAI team has released their openai/circuit-sparsity model on Hugging Face and the openai/circuit_sparsity toolkit on GitHub. The release packages the models and circuits from the paper ‘Weight-sparse transformers have interpretable circuits‘. https://arxiv.org/pdf/2511.13653 What is a weight sparse transformer? The models are GPT-2 style decoder only transformers trained on Python code. Sparsity is not added after…