Koen is available for hire

Koen Deschacht

Verified Expert in Engineering

Machine Learning Developer

Location

Antwerp, Belgium

Toptal Member Since

September 30, 2019

Koen is a passionate software developer with a PhD in machine learning. He's worked for startups and enterprise clients and has experience with a wide range of technologies (machine learning, web apps, web scraping) and languages (Java, JavaScript, Python, Rust). With a love of learning, excellent communication skills, and experience leading and working teams of all sizes, you can be sure Koen will hit the ground running on your project.

Natural Language Processing (NLP)Java Agile Software Development Machine Learning Data Scraping Scrum CSS Git Databases MySQL Android Python Sass Vue Python 3 Beautiful Soup RocksDB

Portfolio

Colstart AI

Vue

Stepwise

JavaScript, Java, Rust, Vue, Regression Modeling

Cegeka

Web3.js, Spring Boot, JavaScript, Java, Angular, Models, Cluster, Node.js...

Experience

Java - 10 years REST - 8 years JavaScript - 6 years Machine Learning - 5 years SQL - 3 years Python - 2 years Angular - 2 years Vue - 2 years

Availability

Part-time

Preferred Environment

PyCharm, CLion, Git, IntelliJ IDEA

The most amazing...

...project I've worked with Microsoft to bring Hyperledger Fabric to the Azure cloud.

Work Experience

Web Developer

2019 - 2019

Colstart AI

Created a website that allowed users to visually browse and manage their data.
Optimized the website for mobile users and visually impaired users.
Implemented an Angular front end that ingested data from a back-end API.

Technologies: Vue

Founder

2018 - 2019

Stepwise

Created a smart digital coach for runners using machine learning in Java.
Built a novel visualization of a runner's progress in Java.
Developed an Android and iOS mobile app with a shared Rust native library and Vue.js user interface.
Performed experiments to measure market demand using Google Ads and Google Analytics.
Conducted interviews with potential customers which resulted in identifying pain points, must-have features, and nice-to-have elements.

Technologies: JavaScript, Java, Rust, Vue, Regression Modeling

Blockchain Developer

2017 - 2018

Cegeka

Built a large number of blockchain proofs-of-concept for Belgian and Dutch banks and insurers. These applications ran on the Ethereum or Hyperledger Fabric blockchain, used a Spring Boot back end and an Angular front end.
Developed an AI algorithm for the automatic reconciliation of payments to assist accountants. The reconciled payments and additions by the accountant were then stored on the Ethereum blockchain.
Created an Xamarin application for Android tablets and iPads with a graphical interface to define a new blockchain application and automatically generate contracts for Ethereum and Hyperledger Fabric.
Assisted Microsoft in the creation of an Azure template to deploy Hyperledger Fabric to the Azure cloud.

Technologies: Web3.js, Spring Boot, JavaScript, Java, Angular, Models, Cluster, Node.js, Ethereum, Hyperledger Fabric

CTO

2015 - 2016

ScriptBook

Designed a scalable and secure software infrastructure. Recruited a team of software developers and machine learning experts. Gave background information on the used technologies to our customers.
Composed multiple machine learning algorithms using Scikit-learn to analyze a movie script, character development, sentiment progression, predicted box office success, and market positioning.
Developed an Angular app that allowed end users to upload a new movie script and to view the results of the automatic analysis.
Created graphs in D3.js to display predicted range of box office earnings, positioning of the movie with respect to other movies, characters in the movie, sentiment, and so on.
Built a crawler for data on movies from public websites using Beautiful Soup and Python.

Technologies: D3.js, Angular, JavaScript, Python, Machine Learning, Scikit-learn

Software Developer

2015 - 2015

NoisyChannels

Built a Java application that would analyze the content of incoming emails using natural language processing and automatically route the email to the correct department.
Constructed a web scraper using Java and Jsoup to extract information from web pages for a large HR company to keep track of legal updates and changes.
Developed an Angular website to view the extracted website information.

Technologies: JavaScript, Angular, Jsoup, Java, Machine Learning

Web Developer

2012 - 2014

Foreach

Developed a public website of a large Belgian newspaper using Java, Spring, and JSP.
Implemented a search engine that indexed more than a million newspaper articles and was updated in real-time using Java and Solr.
Constructed a search results page for a large Belgian restaurant platform using Java and Solr.
Built the back end of a website (Java, Spring) where journalists write their online and newspaper articles and select and edit pictures.
Improved the speed of a slow Umbraco C# website by profiling and removing bottlenecks—achieving an average 10x speedup.

Technologies: jQuery, Solr, Oracle SQL, Thymeleaf, Jakarta Server Pages (JSP), Spring, C#, Java

Lead Developer

2011 - 2012

Cult

Developed a crawler using Java for collecting geolocations.
Built an iOS app that allowed users to chat and share images based on geolocation.
Composed a novel algorithm to determine the relatedness of users based on visited locations and shared friends.

Technologies: Objective-C, Cassandra, Java

PhD and Research Assistant

2005 - 2011

KU Leuven

Used natural language processing in Java to extract names, locations, and actions from fan scripts.
Developed a novel latent model for weakly unsupervised machine learning on texts.
Coached several students in the development of their master theses.

Technologies: MALLET, Stanford CoreNLP, Java

Experience

Count-db

https://github.com/koendeschacht/count-db

Count-db is a very fast key-value store written in Java. It was developed to be used by large scale language models and other machine learning models and outperforms LevelDB, Kyoto Cabinet, and RocksDB.

Terashare

http://terashare.net/

Terashare makes sharing files huge files (gigabytes to terabytes) easy by wrapping the robust peer-to-peer BitTorrent protocol in a very user-friendly interface.

Education

2006 - 2010

PhD Degree in Computer Science

KU Leuven - Leuven, Belgium

2001 - 2005

Master's Degree in Informatics (Computer Science)

KU Leuven - Leuven, Belgium

Skills

Languages

Java, Python 3, JavaScript, Python, CSS, Sass, C#, Objective-C, Rust, Java 10, Bash Script, SQL

Tools

IntelliJ IDEA, Git, CLion, PyCharm, Cluster, Google Analytics, Solr, Stanford CoreNLP

Paradigms

Agile Software Development, Concurrent Programming, Scrum, REST

Other

Natural Language Processing (NLP), GPT, Generative Pre-trained Transformers (GPT), Data Scraping, Machine Learning, Regression Modeling, Models, Scrum Master, Google Ads

Frameworks

Angular, Spring Boot, Spring, Thymeleaf, Jakarta Server Pages (JSP), Apache Struts 2

Libraries/APIs

Vue 2, Vue, Node.js, Web3.js, D3.js, jQuery, Scikit-learn, Jsoup, MALLET, Beautiful Soup

Platforms

Android, Linux, Ethereum, Hyperledger Fabric, Blockchain

Storage

Databases, MySQL, Cassandra, Microsoft SQL Server, Oracle SQL, LevelDB, RocksDB

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring