Software - Albert Alonso

This is a selection of open-source and research-related software I’ve developed. You can also find my GitHub for details.

Highlighted Open-Source Libraries

Bayex: Bayesian Optimization in JAX

Code ReadTheDocs

Simple yet powerful Bayesian Optimization library using JAX and XLA compilation. Ideal for black-box optimization in research workflows.

PCAx: Differentiable PCA in JAX

Code

A lightweight, differentiable PCA implementation in JAX, designed for seamless GPU acceleration. Enables backpropagation through PCA layers with minimal overhead.

t-SNEx: Minimal t-SNE in JAX

Code

High-performance, minimal t-SNE implementation for visualizing high-dimensional data. Built with scalability and GPU efficiency in mind.

BoundVor: Bounded Voronoi Tessellation

Code

Wrapper around scipy.spatial.Voronoi to clip diagrams to arbitrary polygons. Useful for modeling bounded spatial domains in physical systems.

Notata: Scientific Logging

Code ReadTheDocs

A lightweight library for saving simulation results in a standardized, reproducible format. Designed for scientists and researchers carrying experiments.

Highlighted Research Code

De(ep)tangle: High-Density Worm Tracking

Code

State-of-the-art tracking system for overlapping slender bodies in biological images. Published in Nature Communications Biology (2023). → Read the paper

ChemoXRL: Reinforcement Learning for Chemotaxis

Code

Deep RL (PPO) implementation for high-noise environments in cellular navigation. Used in PNAS Nexus (2024) and related works. → Read the paper

Recloc: Receptor Clustering and Sensing Accuracy

Code

Simulation framework and analysis toolkit for studying spatial receptor clustering and its effect on sensing performance. Used in the PRL 2025 paper: Local clustering but global spreading of receptors for optimal sensing. → Read the paper

GradNodes: Adaptive Node in Transport Networks

Code

Differentiable framework to explore optimal node positioning in spatial networks under flow constraints. Part of the work presented in the 2024 preprint on adaptive node placement. → Read the paper