Ayi NEDJIMI - Cybersecurity & AI Tools

100+ open-source cybersecurity, AI, and GPU computing tools

View on GitHub

CUDA/GPU & MLOps Tools

10 GPU-accelerated tools leveraging NVIDIA CUDA for AI/ML workloads

Back to Home


Overview

These tools are built for AI/ML engineers and security researchers who need maximum GPU performance. They leverage CUDA for parallel processing, achieving speeds impossible with CPU-only solutions.


Tool Description GPU Required
CUDAEmbeddings GPU-accelerated embedding server - 10,000+ embeddings/sec Yes
GPUQuantizer Model quantization: FP16/INT8/INT4 with CUDA kernels Yes
VRAMSwapper Intelligent VRAM/RAM offloading for oversized models Yes
ADBloodHound-AI AD attack path analysis with graph neural networks Optional
YaraGen-AI AI-powered YARA rule generation from malware samples Optional
KQLHunter Natural language to KQL query translation No
ModelBench LLM benchmarking: latency, throughput, memory profiling Yes
DatasetForge Cybersecurity dataset generation pipeline No
HashCracker-GPU GPU-accelerated hash analysis (MD5, SHA, NTLM, bcrypt) Yes
PacketSniffer-AI ML-based network traffic classification and anomaly detection Optional

Hardware Recommendations

Component Minimum Recommended
GPU NVIDIA GTX 1080 (8GB VRAM) NVIDIA RTX 3090 (24GB VRAM)
CUDA Toolkit 11.8+ 13.0+
System RAM 16 GB 64 GB
Storage 100 GB SSD 500 GB NVMe
CPU Intel i7 / AMD Ryzen 7 Intel i9 / AMD Ryzen 9

Performance Benchmarks

Tool CPU Performance GPU Performance Speedup
CUDAEmbeddings 200 emb/sec 10,000+ emb/sec 50x
HashCracker-GPU 1M hash/sec 5B hash/sec (NTLM) 5000x
GPUQuantizer 30 min/model 2 min/model 15x

Installation (Common)

# Prerequisites
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu118
pip install nvidia-cuda-runtime-cu12

# Install any tool
git clone https://github.com/ayinedjimi/TOOL-NAME.git
cd TOOL-NAME
pip install -r requirements.txt

Back to Home 2026 Ayi NEDJIMI