01.07.2026 aktualisiert
teilweise verfügbar
info: Abends und am Wochenende verfügbar
AI Engineer
Schwetzingen, Deutschland
Deutschland
MSc Computational LinguisticsSkills
ForschungKünstliche IntelligenzData AnalysisMicrosoft AzureDatenbankenContinuous IntegrationDevOpsGitHubPythonMachine LearningMicrosoft Sql-ServerNatural Language ProcessingNumPyAzure Machine LearningTransformer
Expert in: Natural Language Processing, Azure ML, Data Analytics, Research
- Languages: Python
- Build: Git, Github
- Tools/ Technologies: Scikit-learn, Pandas, Numpy, Matplotlib, Seaborn, PyTorch, Transformers, PyTorch Geometric, Plotly Dash, spaCy, and related ML stack
Advanced in:
- DBMS: MS SQL Server
- DevOps: Azure CI/CD, Azure Key vault
- Deployment and Monitoring: FastAPI, Azure Functions, Streamlit, Weights and Biases, MLflow
- LLM: spacy-llm, LangChain
Certificates:
- AI-102: Azure AI Engineer Associate (Microsoft)
- Deep Learning Specialization (Coursera)
- Applied Data Science with Python Specialization (Coursera)
- Languages: Python
- Build: Git, Github
- Tools/ Technologies: Scikit-learn, Pandas, Numpy, Matplotlib, Seaborn, PyTorch, Transformers, PyTorch Geometric, Plotly Dash, spaCy, and related ML stack
Advanced in:
- DBMS: MS SQL Server
- DevOps: Azure CI/CD, Azure Key vault
- Deployment and Monitoring: FastAPI, Azure Functions, Streamlit, Weights and Biases, MLflow
- LLM: spacy-llm, LangChain
Certificates:
- AI-102: Azure AI Engineer Associate (Microsoft)
- Deep Learning Specialization (Coursera)
- Applied Data Science with Python Specialization (Coursera)
Sprachen
DeutschGrundkenntnisseEnglischverhandlungssicher
Projekthistorie
- Used Generative AI to create a dataset for Biomedical Named Entity Recognition in a low-resource setting
- Implemented a method to create a Biomedical Knowledge Graph using Wikidata for Entity Linking in German
Publications:
- Mustafa, F. E., Dima, C., Diaz Ochoa, J. G., & Staab, S. (2024). Leveraging Wikidata for biomedical entity linking in a low-resource setting: a case study for German
- Ochoa, J. G. D., Mustafa, F. E., Weil, F., Wang, Y., Dima, C., Kama, K., & Knott, M. (2023). The Aluminum Standard: Using Generative Artificial Intelligence Tools to Synthesize and Annotate Non-Structured Patient Data
- Use Graph Neural Network (GNN) in the link prediction setting to exploit potential graph-structured information present in the dataset
- Error analysis and a plausible explanation for the substandard performance achieved by GNN
Publications:
- Mustafa, F., Boutalbi, R., & Iurshina, A. (2023, May). Annotating PubMed Abstracts with MeSH Headings using Graph Neural Network. In Proceedings of the Fourth Workshop on Insights from Negative Results in NLP (pp. 75-81)
- Created graph and initialized nodes with different types of patient embeddings e.g. Autoencoder embedding to reduce the dimensionality of ICD vector
- Implemented Graph Neural Network to recommend GOPs and cluster patients
- Interpreting results using methods like saliency map
- Displaying the results using dashboard
Publications:
- Ochoa, J. G. D., & Mustafa, F. E. (2022). Graph neural network modelling as a potentially effective method for predicting and analyzing procedures based on patients' diagnoses. Artificial Intelligence in Medicine, 131, 102359