Highlighted Projects
Work Experience
Data Scientist at University of Groningen
10/2022 - Present
Collaborated with legal experts and biomedical engineers to develop and apply AI solutions in NLP and computer vision.
- Lead interdisciplinary teams collaborating with legal experts, and biomedical engineers
- Developed and deployed scalable ML pipelines for image and text classification using PyTorch, Docker, DVC, and Flask (check Highlighted Projects)
- Published high-quality research in TPAMI, Neurocomputing, and others
- Conducted AI training sessions for legal experts
- Built a data pipeline using Airflow and BigQuery to automate ETL processes; performed exploratory data analysis (EDA) using SQL (PostgreSQL) and Pandas
Lecturer at University of Groningen
10/2021 - 09/2022
Taught undergraduate and graduate courses on data science and machine learning.
- Delivered lectures on AI concepts and practical ML applications: Neural Networks and Computational Intelligence, Introduction to Data Science, Introduction to Machine Learning, Introduction to Scientific Computing.
- Supervised three bachelor and master students in AI projects
PhD at University of Groningen
07/2017 - 09/2021
Contributed to the international project SUNDIAL (8 universities in 7 countries), and developed end-to-end solutions for analyzing large astronomical datasets.
- Developed novel algorithms for the detection and characterization of astronomical structures
- Co-supervised four bachelor and master students in AI projects
- Provided training and support as Teaching Assistant for AI-related courses
Data scientist Intern at Webfleet Solutions
06/2019 - 07/2021
Worked in a cross-functional team of data engineers and data scientists at Webfleet, a leading telematics provider.
- Developed a predictive model leveraging telematics data to estimate the position of the Link Box, a device utilized for tracking car conditions, within a vehicle
- Worked in an agile environment.
Data Scientist Intern at Thunderbyte AI
05/2020 - 07/2020
Contributed to an international project, called DayTime project, on Digital Lifecycle Twins for predictive maintenance involving Philips and Thundebyte AI.
- Developed NLP models for automated analysis of MRI machine log files, achieved 93% accuracy.
- Enabled faster MRI log data analysis and troubleshooting
AI researcher at Mittweida University of Applied Science
10/2016 - 05/2017
Contributed to the Computational Intelligence Group CI, which focuses on the classification of high-dimensional and Big Data using machine learning methods
- Researched alignment-free sequence learning; published in ICAISC. The method is particularly relevant for shift-invariant sequence comparison and applicable to biological data such as DNA sequences.
Projects
Brain Tumor Classification
A Flask-based web application designed to assist in the diagnosis of brain tumors using MRI scans. It employs a machine learning model to detect three types of brain tumors and categorized into four classes: glioma, meningioma, pituitary tumor, and no tumor.
Key Highlights:
- High Accuracy: It achieves 99.7% accuracy in classifying brain tumors.
- Model Explainability: Incorporates explainability features to help users understand the model's predictions.
- Web Interface: Built with Flask, the application provides a user-friendly interface for uploading MRI images and viewing predictions.
Legal case classification
A Flask-based web application designed to assist legal professionals in analyzing and classifying legal cases. By using machine learning techniques, the app predicts whether a case pertains to housing or eviction based on the textual information provided.
Key Highlights:
- Flask Web Application: Provides a user-friendly interface for uploading legal case documents and viewing classification results.
- Explainable AI: Visualizes the most influential words from the models, enhancing transparency and interpretability of predictions.
