Zaisha

vLLM vs Ollama: Production Serving Throughput and VRAM Benchmarks on Linux

A rigorous performance test evaluating request concurrency, token throughput, and memory offload between vLLM and Ollama on Linux workstations.

Z

Zaisha

Verified 3 min

Mythical Man-Month in the Age of AI: A Linux Workflow

Brooks's Law still holds when AI writes code: coordination cost outpaces parallel output. This guide instruments a Linux repo to measure where AI assistance actually reduces delivery time.

Z

Zaisha

Verified 11 min

Local Image Generation

SIMD Computer Vision on a $2 ESP32-S3 Microcontroller

Use the ESP32-S3's Xtensa LX7 SIMD instructions to accelerate pixel-level operations for computer vision. This guide covers toolchain setup, vectorized kernels, and performance measurement.

Fixing vLLM Tool-Call Accuracy on Linux

Tool-call accuracy in vLLM often degrades due to mismatched sampling parameters and parser configurations. This guide walks through a Linux-based fix using guided decoding and prompt tuning.

Anthropic S-1 Filing: Build a Local SEC EDGAR Parser

The Anthropic S-1 filing is the top Hacker News story with 973 points, making it a practical dataset for building a local SEC EDGAR parser. This guide shows how to fetch, validate, and extract structured text from the filing on Linux.

Building a Production RAG Agent with LangChain, pgvector, and PostgreSQL on Ubuntu

Deploy a high-performance vector search database and connect it to LangChain for accurate, context-aware document question answering.

Local LLM vs Cloud API for Code Explanations: Cost and Speed Compared

A practical comparison analyzing token latency, monthly electricity vs API expenses, and data privacy trade-offs for code analysis.

Fine-Tuning Llama 3 with LoRA and Unsloth on Linux GPUs: Step-by-Step Guide

A practical tutorial on fine-tuning open-source LLMs on custom instruction datasets with 2x speed and minimum VRAM overhead on Linux.

UFW vs IPTables: Rule Simplicity and Firewall Performance Compared

A practical comparison between UFW frontend and raw iptables/nftables on Linux servers, highlighting configuration trade-offs and Docker pitfalls.

Building Fast AI CLI Tools with Rust and Ollama on Ubuntu

Learn how to build lightweight, zero-dependency command line AI tools in Rust that communicate with local Ollama instances over Unix sockets.

Complete Setup of Open WebUI with Docker Compose and NVIDIA Container Toolkit

A complete, production-ready guide to running the ChatGPT-like Open WebUI interface on your self-hosted Linux server with hardware acceleration.

Fish Shell vs Zsh on Linux: Daily Workflow and Performance Compared

An evaluation of Fish Shell versus Zsh for Linux developers: weighing zero-config autosuggestions against POSIX script compatibility.

Z

Zaisha

Verified 3 min

Articles by Zaisha

vLLM vs Ollama: Production Serving Throughput and VRAM Benchmarks on Linux

Mythical Man-Month in the Age of AI: A Linux Workflow

SIMD Computer Vision on a $2 ESP32-S3 Microcontroller

Fixing vLLM Tool-Call Accuracy on Linux

Anthropic S-1 Filing: Build a Local SEC EDGAR Parser

Building a Production RAG Agent with LangChain, pgvector, and PostgreSQL on Ubuntu

Local LLM vs Cloud API for Code Explanations: Cost and Speed Compared

Fine-Tuning Llama 3 with LoRA and Unsloth on Linux GPUs: Step-by-Step Guide

UFW vs IPTables: Rule Simplicity and Firewall Performance Compared

Building Fast AI CLI Tools with Rust and Ollama on Ubuntu

Complete Setup of Open WebUI with Docker Compose and NVIDIA Container Toolkit

Fish Shell vs Zsh on Linux: Daily Workflow and Performance Compared