Keila Moral Figueroa
Research Software Engineer, FAIRmat at HU-Berlin | PhD in Particle Physics, ex-CERN
Berlin, Germany
Actions
Keila Moral Figueroa is a Research Fellow in scientific software development for materials science at Humboldt-Universität zu Berlin, where she works as a software engineer in the FAIRmat collaboration. Her main responsibility is the architectural redesign of pynxtools, an open source library for converting experimental data into the NeXus format.
She previously spent nearly four years developing software for the CMS Collaboration at CERN, working on machine learning pipelines, data analysis frameworks for experimental particle physics, and statistical modelling, in particular the Combine statistical tool. She holds a PhD in particle physics, an MSc in theoretical physics from the University of Edinburgh, and a BSc in physics from the Universidad Complutense de Madrid. Her work centres on the engineering practices that keep scientific and machine learning results reproducible.
Area of Expertise
Topics
It Worked in My Notebook: Engineering Reproducible ML Pipelines
Machine learning pipelines often fail in a way conventional software rarely does: they continue to run but produce different results. A model retrained a month later may score differently although no one changed the code. Typical causes include an updated transitive dependency, a new accelerator driver, an undocumented preprocessing step, or an evaluation that calls a hosted model through a "latest" alias. Because nothing crashes, these discrepancies often remain unnoticed until results must be explained, audited, or reproduced.
This talk treats reproducibility as an engineering discipline. Drawing on practices from research software, where results must remain verifiable for years, it examines four sources of drift: dependencies and hardware, data and preprocessing, numerical nondeterminism, and hosted model versions. For each, it presents a concrete control: data and model versioning, deterministic training configuration, and continuous integration that validates results as well as execution. Examples use JAX, Keras, Colab, and the Gemini API, and the patterns apply to any stack. Attendees will leave with a tiered checklist for their own pipelines and clear criteria for deciding how much reproducibility a given project requires.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top