Ioannis Koumarelas

Ioannis Koumarelas

PhD graduate in Data Cleaning

Hasso Plattner Institute

Mission statement

Great applications can change the world and having high-quality data is more important than we previously thought. I am passionate about understanding data and making them a powerful tool in our hands through sophisticated Machine Learning and Data Engineering solutions.

Interests
  • Data Matching
  • Machine & Deep Learning
  • Natural Language Processing
Education

  • PhD in Information Systems, 2020

    Hasso Plattner Institute

  • MSc in Information Systems, 2014

    Aristotle University of Thessaloniki

  • BSc in Information Systems, 2012

    Aristotle University of Thessaloniki

Skills

Entity Resolution
Data Cleaning
Machine Learning
Data Engineering
Research & Development
Data Science

Experience

 
 
 
 
 
Veeva Link
Data Scientist
Veeva Link
Dec 2021 – Present Remote (Berlin, DE)

Responsibilities include:

  • Duplicate Detection
  • Clustering
  • Creation & Enrichment of medical profiles
 
 
 
 
 
HPI Schul-Cloud
Data Engineer / Machine Learning Engineer
HPI Schul-Cloud
Apr 2020 – Nov 2021 Potsdam, DE

Responsibilities include:

  • Analysis & Presentations
  • Research & Publications
  • Data Cleaning methods
 
 
 
 
 
SAP Concur
Research Consultant
SAP Concur
Nov 2015 – Oct 2018 Potsdam, DE

Responsibilities include:

  • Analysis & Presentations
  • Research & Publications
  • Data Cleaning methods

Accomplish­ments

Udemy
Backend and Frontend technologies
Udemy May 2021
Coursera
Deep Learning Specialization (5 courses)
Coursera Jan 2021
  1. Neural Networks and Deep Learning.
  2. Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization.
  3. Structuring Machine Learning Projects.
  4. Convolutional Neural Networks.
  5. Sequence Models.
See certificate
Udemy
Containers and orchestration
Udemy Dec 2020

Featured Publications

MDedup: Duplicate Detection with Matching Dependencies
Our system uses automatically discovered MDs, various dataset features, and known gold standards to train a model that selects MDs as duplicate detection rules. Once trained, the model can select useful MDs for duplicate detection on any new dataset.
Ioannis Koumarelas, Thorsten Papenbrock, Felix Naumann
PDF Cite Code Video Supplementary code utilities Repeatability
MDedup: Duplicate Detection with Matching Dependencies
Binary Theta-Joins using MapReduce: Efficiency Analysis and Improvements
We deal with binary theta-joins in a MapReduce environment, and we make two contributions. First, we show that the best known algorithm …
Ioannis Koumarelas, Athanasios Naskos, Anastasios Gounaris
PDF Cite Code
Binary Theta-Joins using MapReduce: Efficiency Analysis and Improvements

Recent Publications

Quickly discover relevant content by filtering publications.
Michael Loster, Ioannis Koumarelas, Felix Naumann (2021). Knowledge Transfer for Entity Resolution with Siamese Neural Networks. In ACM JDIQ 2021.

PDF Cite Repeatability

Ioannis Koumarelas, Lan Jiang, Felix Naumann (2020). Data Preparation for Duplicate Detection. In ACM JDIQ 2020.

PDF Cite Code Repeatability Supplementary code utilities

Philipp Schirmer, Thorsten Papenbrock, Ioannis Koumarelas, Felix Naumann (2020). Efficient Discovery of Matching Dependencies. In ACM TODS 2020.

PDF Cite

Ioannis Koumarelas, Axel Kroschk, Felix Naumann (2018). Experience: Enhancing Address Matching with Geocoding and Similarity Measure Selection. In ACM JDIQ 2018.

PDF Cite

Ioannis Koumarelas, Athanasios Naskos, Anastasios Gounaris (2018). Flexible Partitioning for Selective Binary Theta-Joins in a Massively Parallel Setting. In Distrib. Parallel Databases 2018.

PDF Cite Code

See all publications

Popular Topics

Apache Hadoop Apache Spark Classification Data Engineering Data Management Data Matching Data Normalization Data Quality Databases Deep Learning Distributed framework Duplicate Detection Entity Resolution HDFS Knowledge Transfer Learnt Similarity Measure Machine Learning Map-Reduce Matching Dependencies Record Linkage

Contact