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, time series, and vision tasks. At the moment, I make use of smart grid data to support utilities in performing predictive maintenance of electricity grid equipment.
- I am leading data science projects that are based on kedro to enhance structure and collaboration. I warmly recommend the kedro community from which I have learned so much and I try to contribute to kedro community project as much as my time allows.
- 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 have a diverse background spanning engineering, applied math, physics, and high-performance computing.
- I am currently working on kedro-dagster a plugin for orchestrating kedro pipelines with dagster, a modern Python asset-oriented orchestrator.
- 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.