Portfolio

Vertex

Vertex AI

Ray

Ray / Anyscale

Rhoncus Semper

Documentation

rec

36-billion compound virtual screening

DVC

Data Version Control

Cybera

Data Science Fellowship

GSOC

Scientific Software

Magna Nullam

Talks,Publications and Others

About Me

I was born in Bogotá, Colombia, where I earned my B.Sc. in Physics from Universidad Nacional de Colombia. After completing my M.Sc. in Physics at Instituto Balseiro in Argentina, I made my way north to Canada to obtain a Ph.D. in Biophysical Chemistry at the University of Calgary.

My research interests focus on the intersection of computational tools, statistical analysis, and machine learning (ML), especially as they apply to biological systems in medical and pharmaceutical contexts. As a Senior ML Engineer with a background in computational science and a focus on scalable ML systems, I am passionate about automation and performance optimization in complex computational workflows, particularly in the context of drug discovery. I love working with Astral products, specially with uv for managing my Python environments (looking at you torch dependencies! 🌝). I also like to use FastAPI for building scalable and fast ML inference services. 🚀

At Recursion Pharmaceuticals, I built automated training pipelines for chemical property prediction, evaluated and optimized inference platforms like Vertex AI and Anyscale, and contributed to large-scale virtual screening efforts using HPC resources. I also worked as a computational scientist at Cyclica Inc. (later acquired by Recursion), where I developed the company's first ML inference solution to serve internal models with minimal latency and led initiatives to adopt data version control for our curated datasets in ML pipelines. I care deeply about building reliable, maintainable systems and enjoy automating complex workflows with modern tools and infrastructure. Since I'm a big believer in best engineering and documentation practices, I pioneered automated documentation solutions for Recursion's Model Catalogue of chemical property prediction models, introduced API docs for internal tools (using Sphinx!), and implemented unit and integration tests across several ML repositories to improve reliability and maintainability.

Outside of work, I'm a data visualization enthusiast and open-source advocate. When I'm not coding or exploring new tech, you can find me cooking, camping, road-tripping through the Canadian Rockies, or enjoying long bike rides. I'm also improving my sushi-making and baking skills, just in case another pandemic strikes.