Jijun TANG

· Design and implement a data platform architecture providing access to large datasets, data ingestion pipelines and data infrastructure · Explore and evaluate new database technologies to build a scalable, cloud-oriented data platform · Make large and/or complex data more accessible, understandable, and usable by implementing advanced APIs for storage and querying/repartitioning · Create unified data models for data analysis and data exploration, as well as ad-hoc monitoring reporting tools · Interact with different teams to meet their needs (machine learning, content, other engineering teams) · Collaborate with engineering, cloud infrastructure and security teams to better understand their requirements and develop a highly scalable system · Contribute to the improvement of development processes as a member of the Agile development team. Show less

Development of decision support tools: * modelisation of a complex physico-chemical process * Creation, optimization and iterations of several machine learning, deep learning algorithms in Python: - LSTM+ GRU combined with a custom encoder layer of transformer architecture dedicated for time series forecasting * Deployment of algorithms by Flask API * Development environment: Linux + Azure (using GPU NC6) * Exchange with experts Creation of hybrid models: * Digital twin: develope machine learning-based models combined with physicochemical models to predict the end of the catalyst cycle Show less