Lucas ARBABYAZD

Developed a software writing in Python/MATLAB with microservice architecture, practicing ETL, respecting CI/CD through all steps. Create software architecture. Querying the datasets. Integrate all the pre- and post-processing steps on signals: scaling, filtering, down-sampling, binarization, spike detection. Visualize the results with group-comparison, significancy tests. Communication with clients (pharma) on presenting the results; collaborating with R&D to define projects; report to stakeholders.

Collaborate in development of neuroinformatic pipelines in Bash and Python based on Agile methodology, version controlling using Git. Leaded different multi-class and multi-label imbalanced classification problems by creating tree-based benchmark models: Random Forest, XGBoost, CatBoost, LightGBM. (with cross-training) Using Bayesian Optimization techniques for hyperparameter tuning (BOGP, SMAC, TPE) with attention on bias-variance tradeoff. Close collaboration with Data Engineers using Spark and Airflow.

• Create and use advanced supervised/unsupervised Machine Learning algorithms and statistics: regression, scenario analysis, modeling, benchmarking, clustering, random forests, trees, neural networks. • Strong problem-solving, conceptual thinking, understanding of statistics and maths. • Develop, improve, maintain, and verify models for the stated medical purpose. • Good knowledge of collaborative tools (Git) and Agile methodologies.

Modeling and Analysis of spatiotemporal dynamics of patients with Alzheimer's Disease