Stefania Patsou

Senior Backend Software Engineer

• Working for a cross-functional team in agile framework, responsible for implementing a large part of Orfium's product regarding royalties.
• Tech lead role. Conducted technical interviews.
• Coding in Python, using the Django and FastAPI frameworks. Setting up virtualenvs, using poetry for dependency management. Configuring SQLAlchemy and alembic for PostgreSQL.
• Learned how to deploy and create components in AWS using CF. Experience with Lambdas, SQS and Step Functions.
• Learned and implemented GitHub Actions for the CI pipeline of the project and configured Dockerfiles.
• Used factories for testing modules of the project and Datadog for pulling logs and metrics.

Web Developer

• Development in a large scale project of Vodafone as a Backend Developer, cooperating with members of different backgrounds and skills, also mentoring newcomers.
• Coding in Java (using Oracle's ATG Commerce) and Quarkus microservices.
• Good understanding of Agile way of working (Scrum).

Junior Backend Engineer

• An external partner - contractor of Vodafone Greece.
• Coding in Java (using Oracle's ATG Commerce).
• Grasped the business logic behind Vodafone's products (plans), CRM and Siebel systems.

National and Kapodistrian University of Athens

Bioinformatics – Biomedical Data Science

GPA: 9.05

Courses: Machine Learning, Bio-statistics, Image Processing courses, Bioinformatics, Biotechnology, Algorithms in Structural Biology, Biology

National and Kapodistrian University of Athens

Computer Science

GPA: 7.27

Dataflow Representation, Model Inspection and Explainability in ML Pipelines

• This thesis enhances MLInspect by integrating explainability features. It patches methods like LIME, SHAP and others, into a Directed Acyclic Graph (DAG) for visualisation. Additionally, it persists trained models in the DAG, facilitating subsequent inspections. When triggered, a new inspection retrieves the model image from the DAG and executes explainability methods, storing results alongside the model node. This structured approach enables seamless integration of explainability into debugging, enhancing understanding of machine learning models.

See Project

Machine Learning - Neural Networks Exercise

• Implementation in Jupyter Notebook with Python 3+. Apply Neural Networks to MNIST & SVHN datasets.
• Check with plots and metrics, how accuracy of the Neural Network changes depending on how many Hidden Layers exist.

See Project

Bachelor Thesis

• "Street Advisor": an android app using IoT to calculate next position of pedestrian user. It also notifies user for future incident with alerting.
• Programming languages used: Java, Python.
• Server-client setup: clients are the users and server is a PC which receives messages, calculates and sends back information to users.

See Project

Software Development Course

• Sigmod 2018 contest as course project. C implementation with unit testing. Team project.
• Implemented most of the project with the other team members, helping with unit test framework CuTest.
• Can be executed with threads. Radix Hash Join algorithm involved. The target was to improve time applying different number of threads and datasets.

See Project