Alexandros Polyzoidis | Data Scientist and Machine Learning Specialist

Hi, I'm Alexandros Polyzoidis.

A

Skilled in building scalable solutions for e-commerce and logistics. Proficient in Python, SQL, Azure, and deep learning frameworks. Passionate about solving real-world problems through data-driven insights and model deployment.

About

I am a Data Scientist skilled in machine learning, cloud services, and building impactful solutions for e-commerce and logistics. My expertise includes Python, SQL, PowerBI, and Azure tools like Data Factory, Synapse Analytics, and Databricks. I have experience with projects in customer segmentation, credit card fraud detection, and image classification for damaged goods. My work has focused on NLP and computer vision, helping to develop scalable solutions that address real-world challenges.

Quantitative Analysis: Mathematics, Statistics
Programming Languages: Python, R, SQL
Machine Learning & Deep Learning: Scikit-learn, PyTorch, TensorFlow, Keras, etc.
Data Visualization: Tableau, Power BI, ggplot2, matplotlib, seaborn
Cloud & Big Data: Azure ML Studio, Synapse Analytics, Databricks, Data Factory
Tools & Software: GitHub, Jupyter, RStudio, VS Code, PgAdmin, Postman
Soft Skills: Critical Thinking, Communication, Collaboration

Currently, I am working as a freelance data scientist and seeking an opportunity to join a data science or machine learning team where I can contribute, learn, and grow in a dynamic environment.

Experience

Self-Employed

Data Scientist

Developed and implemented customer segmentation models for e-commerce clients, enhancing targeted marketing and personalization strategies.
Created computer vision solutions for supply chain and logistics companies, including image classification to detect damaged loads during loading and unloading processes.
Leveraged Python, Scikit-learn, and Azure ML Studio to build scalable models; utilized Tableau and Matplotlib to present insights effectively.
Tools: Python, Scikit-learn, Tensorflow, Keras, Pytorch, Azure ML Studio, Tableau, Matplotlib, PowerBI

October 2023 - Present | Remote

CYCLOPT

SaaS B2B Sales & Strategy Analyst, Revenue Department

Conducted exploratory data analysis and applied K-means clustering to segment prospective clients and prioritize leads, contributing to an annual revenue growth nearing €400,000.
Utilized Tableau and Power BI to visualize key sales metrics, generate actionable insights, and inform strategic decisions, optimizing outreach efforts.
Leveraged Odoo CRM to manage daily activities, accurately forecast pipelines, and develop data-driven quotes and proposals.
Tools: Tableau, Power BI, Odoo CRM, Python, K-means Clustering

January 2022 - May 2024 | Thessaloniki, GR

SOFTWEB ADAPTIVE SOLUTIONS

Account Manager

Enhanced client relationships by implementing data-driven strategies, resulting in a 20% increase in client satisfaction ratings.
Utilized Entersoft CRM software to track client interactions and identify trends for targeted strategies.
Collaborated with the analytics team to analyze sales data and provide personalized recommendations to clients.
Tools: Entersoft CRM, Data Analytics, Client Relationship Management

May 2021 - January 2022 | Thessaloniki, GR

MAKIOS LOGISTICS

Transportation Manager

Coordinated and planned transport activities for the Packed-Goods Transport Department, increasing on-time delivery rates by 15%.
Minimized empty cargo trips by 20% by strategically reorganizing truck schedules based on customer demand and inventory levels.
Streamlined route planning for 50+ vehicles using GPS and optimization software, boosting delivery efficiency by 25% and cutting fuel use by 15%.
Tools: GPS, Route Optimization Software, Data Analytics, Excel

April 2018 - August 2019 | Thessaloniki, GR

Projects

                
Petra Reviews Sentiment Analysis
                  An end-to-end AI application that classifies visitor reviews into positive, neutral, or negative sentiments.
                
AccomplishmentsTools: FastAPI, PyTorch, Docker, Transformers (Hugging Face), Jinja2, Chart.js, Python
Achieved ~90% accuracy using optimized transformer-based models.
Interactive web interface with sentiment trend visualization.
Deployed using Docker on Heroku for scalability and ease of access.

Endangered Species Classifier
                  A deep learning model to identify endangered animal species based on image classification.
                
AccomplishmentsTools: TensorFlow, Python, Streamlit, Heroku, ResNet50, OpenCV, NumPy, Keras, MLflow, Matplotlib, Docker
Classifies species with high confidence using transfer learning.
Visually appealing web interface with responsive design for ease of use.

Emotion Image Classifier
                  A deep learning project that classifies the emotion (Happy or Sad) in uploaded images.
                
AccomplishmentsTools: TensorFlow, Python, FastAPI, Heroku, Docker, OpenCV, NumPy, Keras, MLflow, Matplotlib,
Achieved high accuracy with MobileNetV3 architecture, optimized for performance.
Supports batch processing and ZIP uploads for efficient classification.

Credit Card Fraud Detection
                  A deep learning model in PyTorch designed to classify fraudulent credit card transactions.
                
AccomplishmentsTools: Python, PyTorch, SMOTE, Scikit-learn, Matplotlib
Trained deep learning model to detect fraud in transactions.
Balanced dataset with SMOTE, achieving high precision and recall.

Card Deck-Image Classification
                  An image classification project that identifies different cards in a deck using PyTorch.
                
AccomplishmentsTools: PyTorch, Python, Image Classification, Computer Vision, Convolutional Neural Networks (CNNs)
Created custom dataset class for efficient card image handling.
Achieved 97% accuracy using EfficientNet-B0 and custom classifier.

Titanic Survival Prediction
                  A predictive model using machine learning techniques to estimate passenger survival likelihood on the Titanic.
                
AccomplishmentsTools: Python, Scikit-learn, Classification, Data Preprocessing, EDA
Predicted Titanic survival using machine learning models.
Achieved 81% accuracy for survival prediction.

Digit Recognizer-LSTM Model
                  A sequence model built with LSTM to classify handwritten digits from the MNIST dataset.
                
AccomplishmentsTools: Scikit-learn, Python, Classification, Data Preprocessing, Exploratory Data Analysis (EDA)
Preprocessed and loaded handwritten digit images efficiently.
Achieved 97.83% test accuracy with an LSTM classifier.

Soil Types Identification Analysis
                  A machine learning model to classify soil types based on physical and chemical soil properties.
                
AccomplishmentsTools: R, KMeans Clustering, ggplot2, factoextra
Cleaned and scaled soil data for accurate k-means clustering.
Identified optimal clusters, visualized using ggplot2 for insights.

Countries Visitation Analysis
                  An analysis project to explore visitation data and trends across different countries.
                
AccomplishmentsMined frequent country visitation patterns with Apriori algorithm in R.
Discovered strong visitation link between Cyprus and Greece.
Tuned parameters to refine travel insights from association rules.

Abalone Age Prediction
                  Uses regression models to predict the age of abalones based on physical measurements from a dataset.
                
AccomplishmentsBuilt a linear regression model to predict abalone age based on physical measurements, achieving strong accuracy through feature selection.
Applied PCA to reduce dimensionality, balancing model interpretability with a slight trade-off in predictive accuracy.

Skills

Quantitative Analysis and Programming Languages

Python

R

MySQL

PostgreSQL

Mathematics

Statistics

Libraries

NumPy

Pandas

OpenCV

Scikit-learn

Matplotlib

Seaborn

Frameworks

Flask

Keras

TensorFlow

PyTorch

FastAPI

Cloud & Big Data

Azure

Databricks

Synapse Analytics

Data Factory

Development & Deployment Tools

Postman

Docker

Heroku

MLflow

Tools & Software

Git

Jupyter

VS Code

PgAdmin

Power BI

Tableau

Soft Skills

Critical Thinking

Communication

Collaboration

Education

                  Hellenic Open University
                
Patra, Greece

                  Degree: Master of Data Science & Machine Learning      
                  

                

                    Relevant Courseworks:
                  
Machine Learning & Data Mining
Deep Learning for Computer Vision
Statistical Data Analysis
Big Data Technologies & Cloud Computing
Natural Language Processing

                  Aristotle University of Thessaloniki
                
Thessaloniki, Greece

                  Degree: M.Sc. in Logistics & Supply Chain Management
                  

                

                    Relevant Courseworks:
                  
Supply Chain Analytics
Inventory & Warehouse Management
Logistics Optimization Techniques
Transportation Planning & Operations
Quantitative Methods in Supply Chain

              
                Aristotle University of Thessaloniki
              
            
Thessaloniki, Greece
Degree: B.Sc. in Finance
Relevant Courseworks:
Financial Management
Corporate Finance
Investment Analysis
Accounting Principles
Economics & Quantitative Methods

              CERTIFICATIONS 
            
Microsoft Certified: Azure Data Scientist Associate – Microsoft, 2024
Python A-Z™: Python for Data Science with Real Exercises! (Udemy)
15 Days of SQL: The Complete SQL Masterclass 2024 (Udemy)
Tableau A-Z: Hands-On Tableau Training for Data Science (Udemy)
R Programming A-Z: R for Data Science with Real Exercises (Udemy)

Contact

+30 69********

apolyzoidis@hotmail.com

github.com/AlexandrosPol

linkedin.com/in/alexandrospolyzoidis