PhD in Computer Science
ED386 - Univeristé Paris Cité 12/2022 – present | Paris, France
Thesis Title: "Patient Representation Learning Using Analogical Reasoning"
Master in NLP (Traitement Automatique des Langues)
Université de Lorraine, IDMC 2020 – 2022 | Nancy, France
Master Thesis Title: "Exploring Analogical Inference in the Healthcare Domain"
Bachelor in Psychology
SWPS University of Social Sciences and Humanities 2016 – 2019 | Warsaw, Poland
Python programming language
• Libraries and Frameworks: PyTorch, scikit-learn, NumPy, pandas
• Data Manipulation and Analysis: Data cleaning, data preprocessing, data visualization (matplotlib, seaborn)
• Machine Learning and AI: Implementing ML algorithms, model training and evaluation, hyperparameter tuning
• Natural Language Processing (NLP): spaCy, NLTK, transformers (Hugging Face)
• Deep Learning: Building and training neural networks, CNNs, RNNs, LSTMs
• Development Tools: Jupyter Notebooks, Anaconda, virtual environments
SQL
• Query Writing and Data Manipulation: SELECT, INSERT, UPDATE, DELETE operations
• Tools and IDEs: SQL Developer, Toad for Oracle Medical Data Intrepretation & Analysis
• Electronic Health Records (EHR): Data extraction, preprocessing, and analysis of EHR data
• Clinical Text Analysis: NLP techniques for analyzing clinical notes
• Data Integration: Combining data from various sources, such as EHRs and laboratory results
• Data Cleaning: Handling missing data, noise reduction, and data normalization specific to medical datasets
• Predictive Modeling: Developing models to predict patient outcomes, readmission rates, disease progression, etc.
• Data Privacy and Security: Ensuring compliance with HIPAA and other regulations, anonymizing patient data
• Collaborative Analysis: Working with healthcare professionals to understand data requirements and interpret results
INRIA PARIS
PhD student at Parisanté Campus & Imagine Institute 12/2022 – present | Paris, France
My research focuses on developing advanced techniques for learning patient representations from both unstructured data (e.g., EHRs, lab results, vital signs, ...) and structured data (e.g., ICD-10 codes, demographics, ...). This involves:
• Experimenting with state-of-the-art models such as BERT, CamemBERT, and RoBERTa to enhance the understanding and processing of medical texts.
• Utilizing various neural network models, including CNNs, RNNs, and LSTMs, to capture complex patterns in the data.
• Applying diverse machine learning algorithms to improve the accuracy and efficiency of patient representation learning.
INRIA PARIS M2
Internship at Parisanté Campus & Imagine Institute 03/2022 – 08/2022 | Paris, France
Key contributions included:
• Gained expertise in EHR systems and developed a convolutional neural network (CNN) based embedding model to learn patient-stay representations from complex medical data.
• Successfully adapted an analogy-based reasoning framework to address two significant biomedical tasks.
• Utilized various NLP techniques for data cleaning and preprocessing.
• Contributed to the field with two papers accepted at: IJCAI - IARML 2022 and ICCBR - ATA 2022 Workshops.
LORIA
M1 Internship at Orpailleur team, LORIA 07/2021 – 08/2021 | Nancy, France
• Enhanced a convolutional neural network (CNN) for a supervised project on "Morphological Analogies," focusing on improving the model's performance for regression task.
• Conducted extensive experimentation to fine-tune the CNN, including adjusting hyperparameters and optimizing learning rates
• Implemented rigorous evaluation techniques to assess model accuracy, robustness, and generalization.
Lionbridge Poland
Localization and Translation Specialist 12/2018 – 07/2020 | Warsaw, Poland
• Translated and localized software applications and content into Arabic, ensuring linguistic and cultural accuracy.
• Conducted quality assurance testing for major projects (e.g., Microsoft, Sony), covering functional, regression, and usability aspects.
• Identified and documented software bugs, collaborating with developers for resolution.
• Developed and executed test plans, cases, and scripts to validate functionality and performance