How to Forge Your Conviction
Strassen's matmul with AVX 512 kernel
Fast Matrix Multiplication with AVX 512 and Loop Tiling
Gumbel Max Trick for Softmax Sampling
3 months of using Neovim
Fail Faster on your Ideas
CUDA Histogram Kernel
Dense Monocular SLAM
Persona Vector Distillation in LLM Weights
Self Doubt Interventions on Chain-of-Thought
Optical Compression
Updating my Bash Prompt
Kubernetes Notes
Paged Attention Performance Analysis
Confidence Aware Router
SSH over Tor
Setting up an SSH Server
LLM Routing Strategies
Median Filter over Arbitrary Datatypes