HNHacker News
TopNewBestAskShowJobs

OsamaJaber

269 karma · joined October 19, 2025

submissionscomments

RunInfra: Optimize any open model down to the kernel, deploy in 5 min

runinfra.ai·2 pts·OsamaJaber·
0

Compiles any HuggingFace model into a single persistent megakernel

twitter.com·2 pts·OsamaJaber·
0

Mega Kernels, Written by Agents

arxiv.org·2 pts·OsamaJaber·
0

AutoMegaKernel: Compiling a LLM into a single CUDA kernel

arxiv.org·3 pts·OsamaJaber·
0

AutoMegaKernel: Compile an LLM into one provably-correct CUDA megakernel

github.com·4 pts·OsamaJaber·
0

StreamIndex: Memory-bounded compressed sparse attention via streaming top-k

arxiv.org·4 pts·OsamaJaber·
0

Show HN: AutoKernel, Auto GPU Kernel Optimization

arxiv.org·2 pts·OsamaJaber·
0

DeepSeek V4's indexer dies at 65K. We got it to 1M on 6GB

arxiv.org·5 pts·OsamaJaber·
0

AutoKernel: Autonomous GPU Kernel Optimization via Iterative Agent-Driven Search

arxiv.org·4 pts·OsamaJaber·
0

DeepSeek V4's indexer OOMs at 65K context. We got it to 1M in 6G

arxiv.org·8 pts·OsamaJaber·
0

Ouroboros: Dynamic Weight Generation for Recursive Transformers

arxiv.org·2 pts·OsamaJaber·
0

Tide: Token-Informed Depth Execution for Per-Token Early Exit in LLM Inference

arxiv.org·3 pts·OsamaJaber·
1

Own your AI. Optimized down to the kernel

runinfra.ai·1 pts·OsamaJaber·
0

Agents with "Hands"

openfang.sh·7 pts·OsamaJaber·
18

Open-Source Agent Operating System

github.com·11 pts·OsamaJaber·
3

PicoLM: Run a 1B parameter LLM on a $10 board

github.com·4 pts·OsamaJaber·
1

The Floating Dock for Developers

snap-dock.co·2 pts·OsamaJaber·
0
← PreviousPage 2 of 2