Hocine Benkelaya

Hocine Benkelaya

Nov 12

Certified kernels, guaranteed at install time

#opensource #machinelearning #python #gpu

2 min read

Hocine Benkelaya

Nov 12

Tokenizing a 12K-token prompt in 0.03 s instead of 25.7 s, and other fixes from running real prompts

#opensource #machinelearning #llm #python

2 min read

Hocine Benkelaya

Nov 12

AI is a neural network, not an API: introducing the NeuroBrix ecosystem

#opensource #ai #machinelearning #python

2 min read

Hocine Benkelaya

Nov 11

Certified kernels instead of autotuning on your machine

#ai #opensource #python #machinelearning

4 min read

Hocine Benkelaya

Nov 10

.nbx: one container format for every AI model

#ai #opensource #python #machinelearning

4 min read

Hocine Benkelaya

Nov 9

Getting the V100 tensor cores back from Triton

#ai #opensource #python #machinelearning

4 min read

Hocine Benkelaya

Nov 8

Build AI apps with Claude Code and NeuroBrix — no PyTorch glue code

#ai #opensource #python #productivity

6 min read

Hocine Benkelaya

Nov 7

nbxtensor: a tensor library for inference that doesn't need PyTorch

#ai #opensource #python #machinelearning

6 min read

Hocine Benkelaya

Nov 7

NeuroBrix 0.6: 49 AI models, one engine, no PyTorch

#ai #opensource #python #machinelearning

7 min read

DEV Community

Badges

Writing Debut

One Year Club

Certified kernels, guaranteed at install time

Tokenizing a 12K-token prompt in 0.03 s instead of 25.7 s, and other fixes from running real prompts

AI is a neural network, not an API: introducing the NeuroBrix ecosystem

Certified kernels instead of autotuning on your machine

.nbx: one container format for every AI model

Getting the V100 tensor cores back from Triton

Build AI apps with Claude Code and NeuroBrix — no PyTorch glue code

nbxtensor: a tensor library for inference that doesn't need PyTorch

NeuroBrix 0.6: 49 AI models, one engine, no PyTorch