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 / collapsible-section-in-gfm.md
Last active September 29, 2025 15:17
collapsible section using direct html elements on top of github markdown flavored with and without code syntax on summary / title

markdown syntax

<details>
<summary> <code> click here to see logs 👇 </code> </summary>

```  
[INFO] [launch]: Default logging verbosity is set to INFO
...
@eevmanu
eevmanu / prompt.txt
Created September 19, 2025 03:45
prompt star framework enhance performance review evidence claim statements
ROLE
You are an expert performance review coach specializing in the STAR framework. Your primary skill is transforming raw notes and statements into compelling, concise, and impactful evidence claims suitable for a formal performance review.
GOAL
Your goal is to help me polish a list of my achievements into well-structured evidence claims using the STAR framework. You will guide me through the process by asking clarifying questions to fill in any missing details and will help consolidate similar points to create a strong, clear narrative for my reviewer.
CONTEXT: THE STAR FRAMEWORK
@eevmanu
eevmanu / prompt.txt
Created September 12, 2025 17:22
prompt to use when need summary on notebooklm and the input is a recording or a video and notebooklm input text limits to around ~300 ish tokens or words
Distill the provided audio file into a detailed technical document for an expert audience.
Meticulously extract and synthesize all significant technical concepts, arguments, evidence, and methodological details from the entire transcript. Preserve the core content, including all nuances, counterarguments, and conclusions, with complete fidelity.
Ensure absolute technical accuracy. Maintain the original's expert-level detail and strictly use its specific terminology and jargon without simplification.
For traceability, substantiate all key claims, evidence, and critical conclusions with brief, targeted quotes or precise references to the source material.
Organize the resulting text logically, reflecting the thematic or argumentative structure of the original discussion.
@eevmanu
eevmanu / 1-prompt
Created August 4, 2025 02:25
deepresearch cli plan spec prd prompt
You are an expert software architect tasked with creating detailed technical specifications for software development projects.
Your specifications will be used as direct input for planning & code generation AI systems, so they must be precise, structured, and comprehensive.
First, carefully review the project request:
<project_request>
ROLE
You are a system architect and AI engineer. Your task is to design a system for a custom, controllable, deep research process, similar in spirit to Grok's "deeper search." The final output should be a conceptual design and pseudocode for a CLI tool that orchestrates this process.
@eevmanu
eevmanu / readme.md
Created August 2, 2025 19:10
DistSys Interview Challenge on twitter x.com - https://x.com/jorandirkgreef/status/1951630005189890266

DistSys Interview Challenge

An infinitely fast, parallel DBMS:

  • executes 2 queries in series,
  • per SQL transaction processed,
  • with 20% transactions updating the same row, 2ms RTT, and no shortcuts on ACID.

Why will this horizontal DBMS not scale beyond N TPS? Solve for N.

@eevmanu
eevmanu / retry-strategies.md
Last active June 25, 2025 15:47
retry strategies - distributed systems - from simplest one to most sophisticated one
  • no retries (the baseline)
    • The caller makes one attempt and propagates any error.
  • simple retry (fixed number of attempts)
    • Retry up to N times as fast as possible.
    • The most basic form of control flow is added: a loop. It introduces the concept of "more than one try" without any timing logic.
  • retry with fixed delay
    • Same as above but waits a constant delay d between attempts.
    • Adds a single, simple parameter—a static wait time. This is the first introduction of temporal decoupling but is otherwise trivial.
  • linear / incremental backoff
  • Delay grows by a fixed increment Δ: t = base + i·Δ.
@eevmanu
eevmanu / 1-code.js
Created June 22, 2025 20:42
simple javascript code to retrieve whole tweet replies from a tweet (working on 20250622)
/**
* This script automates the process of scrolling down a page,
* scraping tweet text, and logging the unique results.
* It's designed to be pasted directly into the browser console.
*/
(async () => {
// --- 1. SETUP ---
// A Set is used to automatically store only unique tweet texts.
const scrapedText = new Set();
const maxScrolls = 100; // Safety limit to prevent an infinite loop.
@eevmanu
eevmanu / 1-prompt.md
Created June 19, 2025 23:09
explanation on how the iterative refinement process works on https://arxiv.org/abs/2505.23060 using gemini-2.5-pro on 20250619

As an expert in self-correcting code generation using large language models, your task is to analyze the following codebase. Your goal is to help me understand its iterative refinement process.

Please follow these steps:

  1. IDENTIFY THE REFINEMENT MECHANISM Explore the codebase and pinpoint the exact functions, classes, or code blocks responsible for self-correction, iterative refinement, or any refinement loop. I need to see where this is explicitly implemented. Please highlight the specific code snippets, including file names and line numbers if possible.

  2. EXPLAIN THE LOGIC Provide a semantic explanation of the code you identified. Infer what the developers are trying to do. Explain step-by-step how the code iterates to refine its self-correcting code generation results or enhance the LLM's output. How does the loop work? What triggers a new iteration? What is the goal of each cycle?

@eevmanu
eevmanu / 1-prompt.md
Created June 19, 2025 22:56
explanation on how the iterative refinement process works on https://arxiv.org/abs/2505.18105 using gemini-2.5-pro on 20250619

As an expert in web-augmented large language models, your task is to analyze the following codebase. Your goal is to help me understand its iterative search and refinement process.

Please follow these steps:

  1. IDENTIFY THE REFINEMENT MECHANISM Explore the codebase and pinpoint the exact functions, classes, or code blocks responsible for self-correction, iterative search, or any refinement loop. I need to see where this is explicitly implemented. Please highlight the specific code snippets, including file names and line numbers if possible.

  2. EXPLAIN THE LOGIC Provide a semantic explanation of the code you identified. Infer what the developers are trying to do. Explain step-by-step how the code iterates to refine its web search results or enhance the LLM's output. How does the loop work? What triggers a new iteration? What is the goal of each cycle?

@eevmanu
eevmanu / python314-t-strings-slides.md
Last active June 17, 2025 02:43
presentation about t-strings, slides content to generate slides using marp (Markdown Presentation Ecosystem) format
theme paginate marp
default
true
true

PEP 750: Template Strings

  • Overview of new t-strings feature in Python 3.14+