It looks like you are in United States. Would you like to go to https://onsiter.com/us/ instead?
Kan være tilgængelig
(Opdateret 2021-03-04)Machine Learning / Data Scientist
København, Denmark
Modersmål Danish, English
- Python
- Natural Language Processing (NLP)
- TensorFlow
Kvalifikationer (18)
XML
NEURAL NETWORKS
DATA SCIENCE
KERAS
Python
Machine Learning
PANDAS
SOFTWARE DEVELOPMENT
Deep Learning
Artificial Intelligence
Image Classification
Natural Language Processing
NLP
Data Analysis
SQL
AI
Git
Kafka
Professionel erfaring
DATA SCIENTIST
GE HEALTHCARE
2020-02 - 2021-02
(1-YEAR CONTRACT)
As a data science consultant for GE Healthcare located at Rigshospitalet (national hospital of Denmark), I work on projects within cardiology. I have been involved in everything from data extraction over preprocessing to designing models and finally
training and evaluating the models.
I have worked mainly in two project areas:
* Machine Learning-based methods for automatic measurements from ultrasounds.
* Machine Learning-based methods for predicting outcome of atrial fibrillation
ablation procedures.
The project related to ultrasound has been focused on automatic measurements of
thickness of interventricular septum (IVS), left ventricular inner dimension (LVID) and thickness of left ventricular posterior wall (LVPW). In particular, I have worked on
comparing the performance of a Machine Learning-based model to the performance of human annotators.
Additionally, as a side project with the hospital, I have worked with myocardial biopsies
from heart transplant patients. The aim of this project is to automatically detect and grade rejection using Machine Learning-based models.
As a data science consultant for GE Healthcare located at Rigshospitalet (national hospital of Denmark), I work on projects within cardiology. I have been involved in everything from data extraction over preprocessing to designing models and finally
training and evaluating the models.
I have worked mainly in two project areas:
* Machine Learning-based methods for automatic measurements from ultrasounds.
* Machine Learning-based methods for predicting outcome of atrial fibrillation
ablation procedures.
The project related to ultrasound has been focused on automatic measurements of
thickness of interventricular septum (IVS), left ventricular inner dimension (LVID) and thickness of left ventricular posterior wall (LVPW). In particular, I have worked on
comparing the performance of a Machine Learning-based model to the performance of human annotators.
Additionally, as a side project with the hospital, I have worked with myocardial biopsies
from heart transplant patients. The aim of this project is to automatically detect and grade rejection using Machine Learning-based models.
LEAD DATA SCIENTIST
RAFFLE.AI
2020-01 - 2020-02
Raffle is AI-powered software that augments customer service flows. I worked on a
Natural Language Processing project aiming to convert natural language text to SQL.
Natural Language Processing project aiming to convert natural language text to SQL.
SENIOR MACHINE LEARNING ENGINEER
KARNOV GROUP
2017-01 - 2020-01
Karnov Group is the leading provider of information to legal and tax professionals in
Denmark and Sweden, and the number one partner in continued legal education in Sweden.
As senior machine learning engineer, I set the vision and direction for data science and machine learning at Karnov. Using machine learning, specifically NLP, I'm building
models for enriching legal documents and helping legal professionals search and find
more easily.
Denmark and Sweden, and the number one partner in continued legal education in Sweden.
As senior machine learning engineer, I set the vision and direction for data science and machine learning at Karnov. Using machine learning, specifically NLP, I'm building
models for enriching legal documents and helping legal professionals search and find
more easily.
CO-FOUNDER AND LEAD DEVELOPER
GUTS & GLORY APS
2014-01 - 2017-01
In 2013 I founded Guts & Glory APS with my girlfriend, Alette Holmberg-Nielsen. We
have a deep desire to create products that the customers really want to use and will
choose to use. We have released an RSS reader for iOS called Readery. It's available on the App Store (unfortunately not yet iOS 12 compatible). We have also built a
recommender system for movies and worked on an Apple TV app for serving users with recommendations. We have also done freelance work, specializing in text-mining
and other data processing tasks. Among our customers were Danske Bank as well as
TNS Gallup in Denmark and Norway.
have a deep desire to create products that the customers really want to use and will
choose to use. We have released an RSS reader for iOS called Readery. It's available on the App Store (unfortunately not yet iOS 12 compatible). We have also built a
recommender system for movies and worked on an Apple TV app for serving users with recommendations. We have also done freelance work, specializing in text-mining
and other data processing tasks. Among our customers were Danske Bank as well as
TNS Gallup in Denmark and Norway.
MACHINE LEARNING ENGINEER
VIVINO APS
2016-01 - 2017-01
Vivino is an app which can identify a wine from a user's photo of that wine. It can then
provide information such as taste notes and ratings.
As a machine learning engineer I was responsible for building Vivino's recommender
system. It's a collaborative filtering system based on implicit feedback, using matrix
factorization.
I've also been responsible for setting up a system for measuring accuracy of the wine
prediction pipeline using Python and R. I have created various prototypes for
improving the pipeline, the best of which have then been put into production. I've also
done various clustering and outlier detection for identifying potential errors in the wine database.
Finally, I have trained image classification models using Keras and techniques such as
deep convolutional siamese networks and deep autoencoders.
provide information such as taste notes and ratings.
As a machine learning engineer I was responsible for building Vivino's recommender
system. It's a collaborative filtering system based on implicit feedback, using matrix
factorization.
I've also been responsible for setting up a system for measuring accuracy of the wine
prediction pipeline using Python and R. I have created various prototypes for
improving the pipeline, the best of which have then been put into production. I've also
done various clustering and outlier detection for identifying potential errors in the wine database.
Finally, I have trained image classification models using Keras and techniques such as
deep convolutional siamese networks and deep autoencoders.
.NET DEVELOPER
GEOMATIC A/S
2012-01 - 2014-01
Geomatic are experts in geographic and demographic data analysis, business
intelligence and segmentation. I was part of a SCRUM team of 5-6 developers. At
Geomatic I primarily developed systems for data analysis. I was working on customer-
facing web applications as well as the machinery for processing data. Most work was
done in C# / .NET and JavaScript / jQuery / Ember for the front end. I have maintained
legacy code in Python as well.
intelligence and segmentation. I was part of a SCRUM team of 5-6 developers. At
Geomatic I primarily developed systems for data analysis. I was working on customer-
facing web applications as well as the machinery for processing data. Most work was
done in C# / .NET and JavaScript / jQuery / Ember for the front end. I have maintained
legacy code in Python as well.
SENIOR APPLICATION DEVELOPER
TNS GALLUP
2008-01 - 2012-01
As senior application developer at TNS Gallup I was responsible for software
development in the Media department. Among many other things I designed and built
a customer-facing web platform for data analysis which is used daily by customers in Denmark as well as abroad.
As responsible for the web platform I handled communication with the sales personnel and in many cases external customers. I worked with customers and talked to them
about their specific needs and afterwards I made the actual implementations as well.
The platform integrates data from a wide range of data sources, for example telephone
interviews, postal surveys and web interviews. The platform is written using C# / .NET and makes use of technologies such as ASP.NET, WinForms and WPF. The front end
relies heavily on AJAX to let the user have a more fluent experience. All data on the
backend is handled by MS SQL Server and each analysis can be configured using XML.
development in the Media department. Among many other things I designed and built
a customer-facing web platform for data analysis which is used daily by customers in Denmark as well as abroad.
As responsible for the web platform I handled communication with the sales personnel and in many cases external customers. I worked with customers and talked to them
about their specific needs and afterwards I made the actual implementations as well.
The platform integrates data from a wide range of data sources, for example telephone
interviews, postal surveys and web interviews. The platform is written using C# / .NET and makes use of technologies such as ASP.NET, WinForms and WPF. The front end
relies heavily on AJAX to let the user have a more fluent experience. All data on the
backend is handled by MS SQL Server and each analysis can be configured using XML.
DEVELOPER
TNS GALLUP
2005-01 - 2008-01
Working as a developer in the Internet Traffic Measurement department and later in the Media department, I have written a wide range of applications for data processing.
This gave me a solid experience in C# / .NET as well as experience improving and extending legacy code in Delphi.
This gave me a solid experience in C# / .NET as well as experience improving and extending legacy code in Delphi.
STUDENT WORKER
NATIONAL CONSUMER AGENCY OF DENMARK, WEB DEPARTMENT
2001-01 - 2004-01
As a student worker it was my job to handle the content of the website and write
smaller JavaScript applications.
smaller JavaScript applications.
Akademisk baggrund
University of Copenhagen and Computer Science, University of Copenhagen
University of Copenhagen and Computer Science, University of Copenhagen
2024-04 - 2024-04
BA
University of Copenhagen
2024-04 - 2024-04
MASTER IN PHILOSOPHY
UNIVERSITY OF COPENHAGEN
2021-03 - 2012-01
BA
UNIVERSITY OF COPENHAGEN
2021-03 - 2009-01
ESPERGÆRDE GYMNASIUM
ESPERGÆRDE GYMNASIUM
2021-03 - 1999-01