Computer Engineering student

Building software.
Exploring AI agents.

I'm Álvaro, a third-year student at the University of Granada. I'm learning how to build software around AI agents, understand their answers, and investigate what goes wrong.

Explore my projects

01 / Selected work

Projects, with context.

What I'm building.
And what I'm learning along the way.

01 Featured project

Learning prototype · In development

git-docs-agent

Git questions. Local notes.
Answers worth examining.

The problem

Can a small assistant answer Git questions from a limited set of notes and recognise when those notes don't contain enough information?

How it works

A Python terminal program prepares five local Git notes and calls Codex CLI. Its instructions ask for answers in Spanish, source citations, and an acknowledgement when the documentation is insufficient. Execution logs help trace what happened in each run.

What I'm learning

Separating program tests from answer evaluation, investigating errors through logs, and refining instructions based on evidence.

  • Python
  • Codex CLI
  • Markdown
  • unittest
Explore git-docs-agent on GitHub
Inside the prototype
  1. 01 / Input A question about Git Asked from the terminal
  2. 02 / Context + tool Five local notes + Codex CLI status · add · commit · diff · push
  3. 03 / Requested response An answer with citations Or an acknowledgement of missing information

Execution logs support investigation.
A manual review examines answer quality.

An early evaluation, with open questions.

The first review found an answer that read the relevant note but omitted required concepts. A revised prompt was tried on two cases; three others still need to be rerun. This limited evidence does not establish general reliability.

Read the evaluation report (Spanish)

02 Building foundations

carrera-tech

Practising the fundamentals,
one useful program at a time.

  • Python
  • JSON
  • Automated tests
Explore carrera-tech on GitHub

The problem

Turning Python study into practical exercises, including a program for recording study sessions and reviewing progress by subject.

How it works

A collection of Python exercises and a study log with input validation, JSON persistence, subject-level statistics, and automated tests.

What I'm learning

Handling invalid input, saving and loading structured data, and using tests to check program behaviour.

02 / A little background

A student.
A habit of asking why.

I'm in my third year of Computer Engineering at the University of Granada. My current interest is software development applied to AI agents: how they use tools and documentation, how their responses are evaluated, and how their errors can be investigated.

I learn through projects, with assistance from Codex, while strengthening my Python, Git, and Linux foundations. For me, building also means taking time to understand the code, check its behaviour, and document its limits.

Education
Computer Engineering · Third yearUniversity of Granada
Foundations in progress
Python, Git & Linux
Also exploring
Introductory AWS training
Languages
Spanish & EnglishNative proficiency in both

03 / Contact

Let's compare notes.

Interested in software, AI agents, or learning by building? Find my projects on GitHub or get in touch on LinkedIn.