About me
I am an Artificial Intelligence (AI) Engineer at Code4Thought. My work includes building trustworthy and ethical AI systems. My research interests include machine learning, ethical AI and information retrieval.
I am passionate about building AI Systems for social good and everything that is data-related.
Download my resumé.
“Those who can imagine anything, can create the impossible.”
Alan Turing
- Machine Learning
- Deep Learning
- Ethical AI
- Data Science
- Information Retrieval
-
Integrated MSc in Computer Engineering, 2016-2022
University of Patras
Skills
Experience
Responsibilities included:
- Data preparation.
- Data visualization with Plotly, Matplotlib and Seaborn libraries.
- Applied Machine Learning algorithms on various problems. More specifically, unsupervised learning (Self Organizing Maps and Gaussian Mixture Models) and supervised learning algorithms (Random Forest, SVM, XGBoost, ADA, Logistic Regression, KNN, CART).
- A deep learning model (U-Net) was built in Keras Tensorflow for a semantic image segmentation task.
My Academic History
Projects
BASH Shell Scripting
Various BASH Shell scripts.
Classification tasks
Red wine quality prediction using Support Vector Machines (SVM) and fake news detection using a Sequential Neural Network with Tensorflow Keras API
Compilers
Designed a compilers' lexical analyzer, parser and BNF Grammar for some-features of Tweet-JSON, using the open-source tools flex and bison.
Data Structures
Interpolation search, binary search, linear search, merge sort, red black tree and digital tree, were implemented in JAVA. Large datasets were used for various observations.
Document Similarity
Top-k most similar documents were retrieved using cosine similarity.
E-recruitment app
A database (MySQL) was designed and a GUI was implemented (in JAVA) for an e-recruitment app. Three types of users are supported, admin, candidate and recruiter.
Mountain Slope
Developed 2 Wordpress plugins & a ski clothing e-shop based on WooCommerce, called “Mountain Slope”.
Multidimensional Data Structures
A B+ Tree was implemented to execute range and exact match queries and a Locality Sensitive Hashing (LSH) algorithm, using the MinHash method for finding similar documents as measured by Jaccard similarity.
Parallel Processing
Parallel implementation of Ramer–Douglas–Peucker algorithm using OpenMP
Patras Gazer
System of collection, management and a crowdsourced information analysis of spatiotemporal information data on human activity using Google Location History, for the city of Patras, Greece.
Scientific computation
Assignments for Scientific Computation course in MATLAB.
Semantic WEB
XML, XML Schema (XSD), XPath and XSL | RDFS ontology | OWL ontology | SPARQL queries
Sentiment classification
Two deep learning models for binary sentiment analysis
Software engineering
Software engineering project (technical deliverables + code).
Thesis
