Hi, I'm Konstantinos – a data scientist and researcher specializing in machine learning and bioinformatics. I like to solve complex biological challenges, manage multidisciplinary projects, collaborate globally, and develop open-source tools.
I am a data scientist and researcher with a strong background in bioinformatics, molecular techniques, and data analysis. My journey began with a degree in Chemical Engineering, where I developed a passion for applying technical problem-solving to real-world challenges. I later pursued a Master’s in Molecular Techniques in Life Science, deepening my expertise in machine learning, data science, and bioinformatics, particularly within the life sciences.
Throughout my academic and professional career, I’ve focused on uncovering biological insights using advanced data-driven techniques. My work includes developing tools like HDAnalyzeR, an R package for proteomics analysis, and contributing to the biomarker discovery field, where I worked in the Human Protein Atlas and identified biomarkers for cancer, metabolic and cardiovascular conditions. These experiences helped me develop my skills in data modeling, visualization, and scientific communication, while allowing me to collaborate with global research teams.
In addition to technical skills, I’ve managed multidisciplinary projects and taken on leadership roles in voluntary organizations, where I learned the importance of teamwork, adaptability, and effective project management.
Looking ahead, I aspire to bridge the gap between science and industry as a biotech or pharma consultant. I’m passionate about helping organizations leverage data-driven solutions to address complex challenges in biology, healthcare, and beyond.
HDAnalyzeR is an R package I developed to streamline the analysis of proteomics data, offering researchers an efficient and user-friendly tool to process and interpret large datasets. This project was driven by the need to enhance the accessibility and reproducibility of proteomic analyses, critical for uncovering biomarkers and understanding disease mechanisms. By designing and implementing this package, I gained valuable experience in R programming, software development, and bioinformatics workflows. I also strengthened my ability to address complex research challenges, optimize data pipelines, and create clear documentation to support the scientific community.
This project focused on identifying biomarker signatures for patients with co-occurring metabolic and cardiovascular conditions using multi-label classification techniques. By analyzing complex proteomics datasets, I managed to classify patients based on overlapping disease states, contributing to a deeper understanding of disease mechanisms and potential therapeutic targets. This experience improved my skills in R programming, advanced statistical modeling, and bioinformatics. I also learned to visualize highly intricate data in an intuitive and efficient manner, enabling clearer communication of complex insights.
SciViz is a Python-based project I initiated to help researchers with an intuitive tool for creating publication-ready visualizations. Recognizing the weakness of Python to visualize complex data effectively and aesthetically, I designed SciViz to simplify and enhance the process of data visualization, making it accessible even for those with limited coding experience. Through this project, I gained hands-on expertise in Python programming, data visualization libraries, and user-focused software development. I also learned to anticipate user needs and create comprehensive documentation to ensure usability.
As part of my Master’s program in Molecular Techniques in Life Science, I undertook a bioinformatics project focused on improving protein secondary structure prediction using neural networks. Leveraging multiple sequence alignment data and a sliding window approach, I developed a model that achieved higher prediction accuracy. To optimize performance, I incorporated advanced techniques such as dropout and ensemble modeling, ensuring robustness and reliability. The project emphasized the critical role of hyperparameter tuning and architectural design in boosting model effectiveness, as demonstrated by strong evaluation metrics.
Dec 2024 - Present
My Master’s thesis focuses on developing a Recurrent Neural Network model constrained by signaling network information to predict time-resolved phosphoproteomic data. This innovative approach aims to advance the understanding of cellular responses to various drugs and doses over time across different cell lines. By applying supervised deep learning techniques, I am working to create a predictive model capable of accurately forecasting phosphoproteomic profiles.
Dec 2023 - Jan 2025
I contributed to the development of the Human Disease Blood Atlas resource, focusing on biomarker discovery through proteomics. I analyzed large datasets and applied machine learning techniques to identify biomarker signatures, as well as developed software solutions for other researchers. This experience enhanced my skills in bioinformatics, data analysis, and collaboration within a multidisciplinary research team.
Jun 2022 - Nov 2022
I worked on the numerical modeling of fine and ultrafine particles in the cardiovascular system, focusing on computational fluid-particle dynamics under the influence of magnetic fields. In this project I developed my skills in computational modeling, data analysis, and problem-solving, while furthering my understanding of complex biological systems through multidisciplinary research.