Skip to content

Instantly share code, notes, and snippets.

View eevmanu's full-sized avatar
🎯
focused

Manuel Solorzano eevmanu

🎯
focused
View GitHub Profile
@eevmanu
eevmanu / claude_mem_search.sh
Created August 4, 2026 21:25
Handoff & Technical Design Specification: Claude Memory Search Script
#!/usr/bin/env bash
# ==============================================================================
# claude_mem_search.sh
# ==============================================================================
# Dynamic interactive search across Claude memory markdown files using fzf + ripgrep.
#
# Features:
# 1. Automatic discovery of memory directories (~/.claude/**/memory/).
# 2. Dynamic ripgrep search triggered only when query length >= 3 characters.
# 3. Instant clean screen when query length < 3 characters.
@eevmanu
eevmanu / pandoc_markitdown_anydoc_comparison.md
Created August 4, 2026 20:27
Comprehensive Technical Comparison: Pandoc vs. MarkItDown vs. AnyDoc

Comprehensive Technical Comparison: Pandoc vs. MarkItDown vs. AnyDoc

Executive Summary

Document conversion is a foundational capability across enterprise search, Knowledge Retrieval-Augmented Generation (RAG) pipelines, publishing systems, and multi-modal AI applications. This authoritative technical report presents a deep, source-code grounded architectural evaluation of three premier document transformation technologies:

  1. Pandoc (v3.10.1): A pure functional document compiler written in Haskell, built around a universal Abstract Syntax Tree (AST) designed for bidirectional multi-target publishing.
  2. MarkItDown: A Python-native object converter pipeline created by Microsoft, designed for flexible, multi-modal ingestion (including LLM vision and audio transcription) targeting Markdown.
  3. AnyDoc (v0.1.3): A ultra-fast, zero-dependency document parsing engine written in modern Rust, engineered specifically by Firecrawl for high-throughput LLM-ready Markdown generation in single-digit m
@eevmanu
eevmanu / steps.md
Created July 31, 2026 17:28
setup to use fff in project with claude code but only local config, don't force it for other peers
which fff-mcp
/home/user/path/to/.../bin/fff-mcp


cd /path/to/project/


jq .enabledMcpjsonServers .claude/settings.local.json
[
@eevmanu
eevmanu / 20260731-git-ignore-exclude-layers-explanation.md
Created July 31, 2026 17:09
understanding different layers of ignore / exclude on git

Confirmed against the official man pages (gitignore(5), git-config(1), gitrepository-layout(5)) and cross-checked against your actual machine — both files exist here and are demonstrably active.

What each path means

$HOME/.config/git/ignore — your per-user, all-repositories ignore list. It isn't hardcoded; it's the default value of the config variable core.excludesFile. Verbatim from git-config(1) (your local 2.54.0 man page):

core.excludesFile Specifies the pathname to the file that contains patterns to describe paths that are not meant to be tracked, in addition to .gitignore (per-directory) and .git/info/exclude. Defaults to $XDG_CONFIG_HOME/git/ignore. If $XDG_CONFIG_HOME is either not set or empty, $HOME/.config/git/ignore is used instead. See gitignore(5).

So the path you asked about is the XDG fallback: XDG_CONFIG_HOME is unset on most Linux desktops, so $XDG_CONFIG_HOME resolves to $HOME/.config, hence $HOME/.config/git/ignore. Same fallback r

@eevmanu
eevmanu / nebius_auth_browser_loop_research.md
Created July 23, 2026 15:08
Technical Research & Implementation Report: Nebius CLI & Kubernetes Exec Plugin OAuth Browser Loop Issue

Technical Research & Implementation Report: Nebius CLI & Kubernetes Exec Plugin OAuth Browser Loop Issue

Document Version: 2.1 (Fully Generalized & Portable Edition)
Target File Path: nebius_auth_browser_loop_research.md
Topics Covered: Nebius Managed Kubernetes (mk8s), OAuth 2.0 PKCE Flow, client-go Exec Plugin Spec, Kubeconfig Hardening (--no-browser, interactiveMode: Never), Dynamic CLI & User Retrieval (which nebius, nebius profile current, jq / grep), Service Account Authorized Keys vs. NEBIUS_IAM_TOKEN Environment Variables.


Executive Summary

@eevmanu
eevmanu / artifact.md
Created July 21, 2026 21:20
The Rise of LLM Inference and AI Systems Engineering: A Deep Dive into Distributed Serving, Architectural Co-design, and Low-Level Optimizations -- LLM Inference Engineering / Inference Systems Engineering -- LLMOps / MLOps

The Rise of LLM Inference and AI Systems Engineering: A Deep Dive into Distributed Serving, Architectural Co-design, and Low-Level Optimizations


Executive Summary

The transition of artificial intelligence from exploratory training paradigms to high-throughput, low-latency, and cost-efficient production serving has catalyzed the birth of two highly specialized, deeply technical engineering disciplines: AI/ML Systems Engineering and LLM Inference Engineering (often grouped under Inference Systems Engineering).

Historically, machine learning was cleanly partitioned: ML researchers designed model architectures in Python, while software engineers deployed them as monolithic black-box containers behind simple HTTP/gRPC gateways. Today, as Large Language Models (LLMs) scale to hundreds of billions of parameters, and as autoregressive generation imposes severe memory, bandwidth, and compute constraints, this clean separation has collapsed.

@eevmanu
eevmanu / file1.md
Created July 21, 2026 18:58
water quality research how to improve in home calidad del agua

Maximizing Residential Water Quality: The Engineering and Biophysical Foundation of the Multi-Stage Gold Standard Purification System

Author: Senior Water Quality Research Specialist
Prepared For: Executive Research Lead
Date: July 21, 2026


Executive Summary

Maximizing residential water quality to clinical and toxicological "gold standards" requires a multi-stage, barrier-in-series treatment train. No single filtration technology is thermodynamic or kinetic-capable of removing the complete spectrum of modern tap water contaminants—ranging from large suspended particulates to sub-nanometer dissolved heavy metals, synthetic halogenated organics, and microbiological pathogens.

@eevmanu
eevmanu / artifact-1.md
Created July 21, 2026 17:44
research about sunscreen

The Sunscreen Dichotomy: Demystifying Mineral Barriers, Chemical Absorption, and the Scientific "Gold Standard"

A Rigorous, Evidence-Based Investigation into Inorganic Photoprotection and Clinical Formulation Safety


1. Executive Summary & The Verdict

The search for the "holy grail" sunscreen is often guided by a mixture of consumer-focused marketing, dermatological advice, and environmental concerns. Consumers frequently seek out "physical" or "mineral" sunscreens under the belief that they act as inert, mirror-like shields that reflect solar radiation away from the skin.

This academic deep-research report synthesizes peer-reviewed literature from cognitive dermatology, photobiology, and clinical pharmacology to evaluate this paradigm and identify the true, scientifically backed "gold standard" in UV protection.

@eevmanu
eevmanu / readme.md
Last active July 19, 2026 16:34
boya castmic p60 docs

https://www.boyamic.com/product/boyacastmicp60

BOYA CastMic P60 - product page

Premium cardioid capsule with 48 kHz / 24-bit audio for studio clarity.
USB plug-and-play connects to phones, tablets, and laptops; XLR outputs to mixers or audio interfaces for uncompromised sound.
Individual mic and headphone volume controls let you adjust input and balance audio mix instantly.
Easily customize EQ, RGB effects, and limiter settings via the BOYA Central app.
90 dB SNR and 130 dB max SPL delivers natural sound with exceptional clarity.
Tap-to-mute provides instant noise control during live streaming and recording.
@eevmanu
eevmanu / readme.md
Created June 29, 2026 02:22
prompt to analize llm inference server math for any model in actual (20260628) gpu server enterprise and consumer formats specs

Universal LLM Inference Mathematics & Co-Design Protocol

A Replication Guide for Future Mathematical Analyses of LLM Serving Bottlenecks

This document is a Universal Replication Protocol. You can copy-paste the template in Section 2 directly into any future AI session to instruct an agent (such as Antigravity) to perform the exact physical and mathematical analysis we developed here for any new model, precision format, or GPU configuration.


1. Quick Reference: Pre-Computed GPU Specification profiles

To make your future runs faster, here are the pre-computed physical parameters for the top enterprise and legacy/mid-tier GPUs (dense FP16/BF16 precision):