Lukas Pfannschmidt is an aspiring researcher working in the intersection between bioinformatics and machine learning. He is focused on methods to automatically uncover relevant elements of biomedical datasets to improve the analysis for other researchers in these areas.

At the moment he is working as part of the CITEC machine learning group to employ novel machine learning algorithms to interesting research data. He is also part of the DiDy international research training group which allowed him to take part in a half-year research exchange at the Simon Fraser University in Vancouver, Canada in 2018.


  • Feature Selection
  • Causal Relevance
  • Biomarker Discovery
  • Parallel Processing


  • MSc in Bioinformatics and Genome Research, 2016

    Bielefeld University

  • BSc in Bioinformatics and Genome Research, 2014

    Bielefeld University

Recent Posts

Reproducible Experiments in Machine Learning

This post describes how we used Docker containers and Python environments to make tables and figures in our recent scientific paper …

Decentralized Website

This article explains how to turn a personal website into a decentralized site based on IPFS, Hugo and GitHub Actions.

FRI Quickstart Guide

Quick start guide In this guide i am going describe how to use the FRI python library to analyse arbitrary datasets. (This guide is a …


Things I am working on right now or which I completed in my studies.

Feature Relevance Analysis Tool

Python tool to discover relevant features in machine learning problems.

K-means OpenCL

Kmeans clustering algorithm in Java utilizing GPU computing.

Adverse Drug Effect Checker

App to display known adverse drug reactions for medical practitioners.

Sublimator Controller

Program to control sublimation machine using Raspberry Pi.

Efficient Edit Distance

Efficient and parallelized implementation of Smith Waterman edit distance algorithm.

Recent & Upcoming Talks

FRI - Feature Relevance Intervals for Interpretable and Interactive Data Exploration

Most existing feature selection methods are insufficient for analytic purposes as soon as high dimensional data or redundant sensor …