I am a ML scientist with 6y+ experience and a passion for ML package development. I enjoy finding simple, interpretable solutions to problems and coding them in a modular way. I believe in FOSS and I look forward to making contributions.
- I have been working on a number of tabular and time series tasks. At the moment, I make use of smart grid data to support utilities in performing predictive maintenance of electricity grid equipment.
- I am enthusiastic about the philosophy and simplicity of the scikit-learn API. I have worked on a number of packages extending and/or integrating with it, most of them unfortunately closed source.
- In the past couple years, I have been excited about uncertainty quantification and especially conformal prediction!
- I am a nerd when it comes to MLOps and I enjoy playing with all the
toystools available. Talk to me about how to structure a data science project so it is production-ready from the get-go or about asset-driven orchestration! - I have a diverse background spanning engineering, applied math, physics, and high-performance computing.
- During my PhD, I studied turbulence models and evaluated them numerically using metaLBM, a C++ simulation package running on GPU clusters using MPI, OpenMP, and CUDA.
- During my postdoc at EPFL, I created giotto-tda, an open-source Topological Data Analysis library for feature engineering and unsupervised learning extending scikit-learn.