Zyphra
Ali Ahmed
I work on high scale back end, data and AI platforms.
65
154
16
1
Projects
Repositories this builder owns.
docker-slurm-llm
Toy distributed llm training with slurm and pytorch
nanoptx
Python Library generation PTX directly with cuda
Aule-Attention
High-performance FlashAttention-2 for AMD, Intel, and Apple GPUs. Drop-in replacement for PyTorch SDPA. Triton backend for ROCm (MI300X, RDNA3), Vulkan backend for consumer GPUs. No CUDA required.
neon-benchmark
Simple project to test neon performance on M1 systems
Connected narratives
Related builders
Others building in the same ecosystem.
