Skip to content

Nataly Jaya Salazar

Assembling

Computer Engineering student · Universitat de Lleida

Nataly Jaya
Salazar

An engineer with a research mindset.

Somewhere between engineering and neuroscience, I'm exploring how computation can help us understand the human brain. Still building and still questioning.

Selected work Download CV

§ 01/About

The ghost behind the shell

Portrait of Nataly at the 4YFN event in Barcelona, in front of curved neon light strips.
Fig. 1 — UPC, Barcelona, 2026

I'm Nataly, a Computer Engineering student at the Universitat de Lleida

I've always been drawn to understanding how things work and what can be built with technology. My background is rooted in software engineering, but my curiosity has been leading me towards machine learning, healthcare, and computational neuroscience. I'm interested in exploring how computation can help us understand something as complex as the human brain.

Education

  • BSc Computer Engineering

    Universitat de Lleida

    2023 — present

  • BachiBac, Science & Technology track

    Vedruna Escorial Vic — French–Spanish dual diploma

    2021 — 2023

Certificates

  • Intro to Deep Learning

    Kaggle

    Oct 2026

  • Intro to Machine Learning

    Kaggle

    Oct 2026

  • Fundamentals of Machine Learning for Healthcare

    Stanford Online

    Sep 2026

  • Python Programming for Solving Mathematical, Physical and Chemical Problems

    Universitat de Lleida

    Sep 2026

Community

  • Frontend developer

    LleidaHack — Pages, components and UI fixes for the student tech association

    2024 — present

  • Marketing member

    LleidaHack — Event visuals, social media and outreach

    2024 — present

Languages

  • SpanishNative
  • CatalanNative
  • EnglishB2
  • FrenchB2

§ 02/Selected work

Three projects, in brief

What each one does and the part I built myself. Violet marks my work.

JetFlow architectureThe IntelliJ plugin pushes code and follows the GitHub Actions run, and triggers deploys on Render and reads their status. The tool window, the Rocky assistant and the Render integration are Nataly's work.IntelliJtool windowRockyGitHub Actionspush → CI runRenderdeploy · status

01/HackUPC 2026

JetFlow

CI/CD without leaving the IDE.

IntelliJ plugin that generates GitHub Actions workflows, follows each CI run and deploys to Render, all from the editor.

My part · team of 3

  • Render deploy layer: the API client and keychain-stored credentials.
  • Deployment tool window and the first Rocky assistant, wired to CI feedback.
  • Kotlin
  • IntelliJ SDK
  • GitHub Actions
  • Render API
StreamSync overviewStreaming platforms feed one catalog, which serves each user's personal library and the staff dashboards. The catalog and the personal library are Nataly's work.platformsCatalogsearch · filtersLibraryyour listsDashboardsfor staff

02/Course project · 2026

StreamSync

One catalog for everything you stream.

Django app that gathers films and series from many platforms into one catalog, with personal libraries and staff dashboards.

My part · team of 4 · most commits

  • Catalog and search, from the external API to the internal database.
  • Onboarding, personal library, settings and notifications.
  • Code-quality metrics in CI, plus issue and PR templates.
  • Django
  • Tailwind CSS
  • Docker
  • GitHub Actions

03/UAB The Hack 2024

Translator

Winner · Rookies

Let your code overcome language barriers.

pip package that translates a Python file's comments, docs and variable names, detecting the source language automatically.

My part · team of 3

  • The translation module: deep-translator's Google backend with automatic language detection.
  • Python
  • deep-translator
  • Poetry
  • PyPI

§ 03/Learning resources

Open notebooks

These are samples and templates I wrote so other people (and future me) can pick up a concept by running it.

  1. Notebook 01

    Data Preprocessing Sample

    Turning raw health data into a model-ready dataset

    A step-by-step notebook on the Life Expectancy (WHO) dataset, written so a beginner can follow every decision before any model is trained. The repo also collects my exercises from the Universitat de Lleida course on Python for mathematical, physical and chemical problems.

    • Sanity checks & EDA
    • Median + KNN imputation
    • IQR outlier capping
    • One-hot encoding
    • pandas · seaborn · scikit-learn
    Open the notebook

  2. Notebook 02

    Multi-Agent System Demo

    Hierarchical LLM agents with LangGraph

    A small, readable demo of the "supervisor of supervisors" pattern. A top-level supervisor routes work between a research team (web search, scraping) and a writing team (outlining, drafting, charts), each with its own supervisor.

    Work in progress — critic, teacher, Obsidian and LaTeX agents are planned

    • LangGraph StateGraph
    • Structured-output routing
    • Tool-using agents
    • Groq · Tavily
    Open the notebook

§ 04/Technical skills

A small connectome

The tools I've used, grouped into regions and wired to their neighbours. No self-rated percentages: each skill points to the project, course or role it comes from.

Languages9

  • Python

    Used in: StreamSync · Translator UABTheHack · Data Preprocessing Sample · Multi-Agent System Demo

  • Kotlin

    Used in: JetFlow

  • JavaScript

    Used in: StreamSync · LleidaHack website

  • TypeScript

    Used in: This portfolio

  • Java

    Used in: University coursework

  • C++

    Used in: University coursework

  • C#

    Used in: LauzHack 2024 (Unity)

  • SQL

    Used in: University coursework

  • Bash

    Used in: University coursework

Web & Backend6

  • Django

    Used in: StreamSync

  • HTML & CSS

    Used in: StreamSync · LleidaHack website · This portfolio

  • Tailwind CSS

    Used in: StreamSync · This portfolio

  • React

    Used in: This portfolio

  • Next.js

    Used in: This portfolio

  • REST APIs

    Used in: JetFlow (Render, GitHub) · StreamSync (catalog API)

AI & Data9

  • pandas

    Used in: Data Preprocessing Sample · UdL scientific Python course

  • NumPy

    Used in: Data Preprocessing Sample · UdL scientific Python course

  • scikit-learn

    Used in: Data Preprocessing Sample

  • Matplotlib & Seaborn

    Used in: Data Preprocessing Sample · UdL scientific Python course

  • Jupyter

    Used in: Data Preprocessing Sample · Multi-Agent System Demo

  • LangGraph

    Used in: Multi-Agent System Demo

  • SciPylearning

    Learning through: UdL scientific Python course

  • Machine learninglearning

    Learning through: Stanford Online — ML for Healthcare · Kaggle — Intro to ML

  • Deep learninglearning

    Learning through: Kaggle — Intro to Deep Learning

Infrastructure & Tools12

  • Git & GitHub

    Used in: Every project here

  • GitHub Actions

    Used in: StreamSync quality gates · This portfolio's CI/CD

  • Docker

    Used in: StreamSync

  • Render

    Used in: JetFlow · StreamSync

  • IntelliJ Platform SDK

    Used in: JetFlow

  • Gradle

    Used in: JetFlow

  • Cloudinary

    Used in: StreamSync

  • Unity

    Used in: LauzHack 2024 · LAN Party Lleida

  • Linux

    Used in: Development environment (CV)

  • Jira

    Used in: LleidaHack

  • Figma

    Used in: LleidaHack

  • LaTeX

    Used in: Reports & CV

§ 05/Research interests & writing

Questions I'm moving towards

I haven't published research. These are the fields I'm reading into, and the place where I'll start writing about them.

  1. I

    Computational neuroscience

    Using mathematical models and simulation to understand how neurons and circuits compute.

  2. II

    Machine learning for healthcare

    How models are built, evaluated and deployed on clinical data — the subject of my Stanford Online course.

  3. III

    Biomedical signal processing

    Turning EEG, ECG and other physiological recordings into features a model can learn from.

  4. IV

    Neurotechnology

    Interfaces between software and the nervous system, and what they could mean for neurological disorders.

Writing — first pieces in progress

  • Substack

    In progress

    First essay — working title to come

    My first long-form piece is being drafted. It will go out on Substack first and be cross-posted to Medium.

    Date
    —
    Reading
    —
    Link
    Soon

  • Medium

    Planned

    Learning notes — coming soon

    Shorter notes from the courses and papers I'm working through, written up as I go.

    Date
    —
    Reading
    —
    Link
    Soon
Follow on Substack@natakenFollow on Medium@njayasalazar

§ 06/Personal journey

Can't Catch Me Now

Conferences and hackathons where I learned to build with other people, in the order they happened. Open one to see its frames.

  1. 2024

    1. 2024/Bellaterra

      My first hackathon, at the Universitat Autònoma de Barcelona: a pip package that translates the comments and identifiers in Python code.

      Developer · Winner, Rookies

      See the project

    2. 2024 — present/Lleida

      The student tech association: Lleidahack team building activity

      Frontend & marketing

    3. 2024/Lleida

      On the organising side with LleidaHack: marketing and activity volunteering.

      Volunteer

    4. 2024/Lausanne

      In Lausanne, Switzerland: a multiplayer drawing game in Unity using augmented-reality glasses and an AR pen.

      Developer

  2. 2025

    1. 2025/Lleida

      Cartes del Mite, a role-playing game built on Catalan myths and traditions.

      Developer · game jam

      Play it on itch.io

  3. 2026

    1. Mar 2026/Barcelona

      4YFN & Mobile World Congress

      Mobile World Congress's startup event, attended as a Universitat de Lleida student.

      Attendee

    2. Mar 2026/Madrid

      A tech conference in Madrid, hosted at the UC3M. The talk on screen: “Hack Your Way to Silicon Valley”.

      Attendee

    3. 2026/Barcelona

      Built JetFlow, a CI/CD plugin for IntelliJ, for the event's IntelliJ challenge.

      Developer

§ 07/Contact

Let's talk about software, data and the brain.

I'm open to internships and research-oriented opportunities in software engineering, machine learning and health technology. Email is the best way to reach me.

njayasalazar@gmail.com
Nataly dabbing beside a life-size R2-D2 in a dark room.
Fig. 7I did the dab.