Pavol Mulinka
Machine Learning Engineer at Forescout(freelancer)
l'Hospitalet de Llobregat, Catalonia, Spain
Summary
🤩
Rockstar🎓
Top SchoolPavol Mulinka is a data scientist with 8 years of experience bridging AI/ML research and production, currently working on healthcare analytics from Catalonia. He started as a network engineer and brings deep systems thinking to ML problems—applying unsupervised and stream-based methods to detect rare events in projects like FIREMAN and analyzing darknet traces at Japan’s National Institute of Informatics. Pavol has practical experience deploying multimodal deep-learning tooling (contributing hyperparameter tuning, RayTune integration and a feature-smoothing layer to a pytorch-widedeep repo) and has applied NLP, LLM prompt engineering and agentic AI to security data at Forescout. He pairs PhD-level telecommunications expertise with hands-on distributed data engineering (PySpark, Kafka, Elasticsearch) and a track record in personalization for mobile games, making him adept at turning complex data into actionable models.
8 years of coding experience
15 years of employment as a software developer
Ing Telecommunications, Ing Telecommunications at Slovenská technická univerzita v BratislaveDoctor of Philosophy (PhD) Telecommunications Engineering, Doctor of Philosophy (PhD) Telecommunications Engineering at Czech Technical University in Prague
Slovak, English, Czech, Spanish
Github Skills (10)
tensorboard10
hyperparameter-tuning10
pytorch10
machine-learning10
deep-learning10
python10
tabular9
datatable9
nlp7
image-processing6
Programming languages (5)
Github contributions (5)
jrzaurin/pytorch-widedeep
Oct 2021 - Jan 2023
A flexible package for multimodal-deep-learning to combine tabular data with text and images using Wide and Deep models in Pytorch
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:1 release, 7 reviews, 133 commits in 1 year 3 months
Contributions summary:Pavol's contributions centered on hyperparameter tuning for a multimodal deep learning model in the context of the pytorch-widedeep framework. They implemented a notebook dedicated to hyperparameter tuning and model visualization using RayTune and Tensorboard for the Protein Homology Dataset. The user also added a RayTuneReporter callback for reporting history and lr_history values and also added a Feature Distribution Smoothing Layer.
multimodal-deep-learningpytorchtabularpythondeep-learning
5uperpalo/FIREMAN-project
Mar 2020 - Jan 2023
Contributions:67 commits, 59 pushes, 1 branch in 2 years 11 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.