Hi, I’m Elisa
A world of opportunity 👋
I apply machine learning methods to solve challenges in biomedicine.
Currently working at Keysight Technologies as an ML researcher, my background includes a blend of machine (and deep) learning modelling & bioinformatics.
Some highlights of the last years:
- Contributed to DeepChem integrating a protein language model to predict binding sites as part of the Google Summer of Code fellowship.
- Recently completed my PhD (yay!), focusing on statistical, ML, and graph representation learning methods to identify blood-based molecular markers for Parkinson’s disease diagnosis and prognosis. Feature selection and engineering, confounding factors & explainable AI. Check it out here!
- During my PhD I got a Fondation de Luxembourg’s fellowship to visit Prof. Pietro Liò (Cambridge University) and Petar Veličković (Google DeepMind). I focused on graph representation learning models for omics data.
- While at Cambridge, racing with the triathlon club the Varsity duathlon against Oxford - quite an experience! Obviously, we won
- Fellow of the 2023 MIT Catalyst program, exploring unmet needs in health(-tech)
- Proud global shaper recognized by the World Economic Forum of Geneva, to make the world a better place.
- Discovered the joy indulging in a good book with the occasional slice of carrot cake or a glass of white wine.
- Some people read my posts on both technical and non technical stuff.
Get in touch if you’re tackling complex ML problems. I’m available for consulting and collaboration 🤗
Disclaimer: this site is currently under (active) construction!
Download my resumé.
- Machine learning
- Graph representation learning
- Omics data
- Drug discovery
-
PhD in Biomedical Data Science, 2024
University of Luxembourg
-
MSc in Bioinformatics, 2018
Universitat Autonoma de Barcelona
-
BSc in Biotechnology (computational), 2017
Universidad Politécnica de Madrid
Skills
Experience
Graph representation learning for modelling omics data (extension of PhD project).
Boosting Prot2text, a deep learning model to predict protein function from sequence and structure data.
I’m part of the Biomedical Data Science group. Integrative machine learning methods for the joint analysis of different types of omics, clinical and imaging data.
Responsibilities include:
- State of the Art research and definition of experiments and project strategy
- Scripting and modelling
- Validation of results
Applied statistics and ML models in a wide range of industries and projects. Forecasting sales, computer vision OCR, statistics for fraud detection, visualization.
Responsibilities include:
- Extraction, preprocessing and analysing data from different formats
- Modelling
- Preparing scripts and models for deployment
Use of SAS to access, manage, analyze and present data. Daily tasks related to SAS developer role, consultant and a migration of data from different database versions (data warehouse).
Responsibilities include:
- Data extraction on SAS
- Database migration
- Scripting on SAS/SAS-SQL
Heuristics for computing hydrophobic properties based on quantum mechanics calculations applied to virtual screening and the alignment of molecules in drug discovery. Built a pipeline x30 times faster without compromising accuracy.
Responsibilities include:
- Research in virtual screening techniques
- Parametrization of hydrophobic descriptors
- Build and validate the parametrized pipeline on external data
Accomplishments
Recent Posts
Side Projects
SleepGuard
Objective sleep data to guide treatment of sleep disturbance in PTSD. Project from MIT-Catalyst program.
Farmelody
Data-driven ‘farm-tech’, our secret weapon is the analysis of microbiome data. Prototype under construction!