Pau Martínez i Ríos

Computer Science student at the Universitat Politècnica de Catalunya. Highly motivated in seeking new opportunities to learn and grow.

Highlighted Projects

Some of the academic or personal projects related to computer science that I have had the opportunity to contribute to.

CliffyJump

3D platformer built in Unity where the player advances through multiple levels by timing jumps across moving and static platforms. Emphasis on tight controls, readable level design, and progressive difficulty.

Implemented modular level prefabs, obstacle logic, and camera framing for clear navigation and jump timing.

Unity
C#

Speaking Bot (IELTS)

AI-powered speaking practice assistant focused on IELTS oral exams. Simulates interview flow, provides prompts, tracks speaking time, and delivers feedback on fluency, coherence, and vocabulary range.

Born from a personal need while preparing for the exam, designed to offer an on-demand practice partner that adapts to user responses.

JavaScript
Laravel
Speech-to-Text / Text-to-Speech
LLM API

Animated Series Classifier

Computer vision pipeline that takes a screenshot from an animated series, extracts visual features, and predicts the most likely series. Explores transfer learning and classical feature descriptors for robust classification.

MATLAB

Generative Models: Cats & Dogs

Training and comparison of generative models on a small cats vs. dogs dataset to synthesize new images. Evaluated architectures and regularization strategies, focusing on sample quality and training stability.

Includes reproducible experiments, data augmentation pipeline, and image grids for visual inspection of model outputs.

Python
PyTorch
GANs / VAEs
NumPy

SQL Interpreter

Created an interpreter for SQL using Python and ANTLR4. Integrated Pandas to handle CSV data and Streamlit for a simple graphical interface.

Python
ANTLR4
Pandas
Streamlit

Implementation of a Data Structure

Implementation of a KD-Tree to organize points in k-dimensional space and enable efficient nearest-neighbor searches via hierarchical space partitioning.

Use of Artificial Intelligence Algorithms

Applied local search algorithms (Hill Climbing and Simulated Annealing) to optimize bike redistribution planning for urban stations, minimizing transport costs.

Java
Local search