Zach B.

Big data analysis of the electrolyser production line at Genvia ( CEA & Schlumberger funded). Detection of anomalies and search for correlation with production data, data cleaning and storage, numerical modeling. Familiarisation with Industry 4.0, manipulation of time series, various physical quantities and visual data. Handling of Python, Databricks, Jupyter Notebook, Pandas, PySpark, OpenCV, Apache Parquet, InfluxDB.

Collaboration with IBM research Zurich for the ontology based (UMLS) enrichment and exploitation of knowledge graphs with biomedical and biochemical scientific literature. Development of deep learning algorithms for the extraction of entities and semantic relations from textual data by supervised, few-shot, zero-shot and distant learning. Implementation on millions of documents. Handling of Python, PyTorch, Hugging Face (GPT-J, T5, BERT), Deepspeed, Deep Search, spaCy, Spark NLP, Docker

Supervised detection of weather phenomena in numerical simulation models. Use of RNN and CNN networks. Work with geospatial, temporal, textual and visual data. Production of the results obtained on real time data and visualisation using dashboards. Implementation of different hyperparameter optimization strategies to go from 0.7 to 0.81 in model accuracy. Handling of Tensorflow, DeepLab2, xarray, cartopy, Horovod, conda, Voilà, Slurm.