Chief Data Scientist · Copenhagen
Esben Eickhardt, PhD
I build machine learning that runs in production. I test assumptions against data instead of guessing, keep solutions as simple as the problem allows, and when something is missing or broken, I build the fix. Evenings go to self-hosted apps, games and maps on my own server.
- 9 yrs
- as a chief data scientist
- PhD
- bioinformatics, Aarhus University
- 36
- repositories on GitHub
- 8
- published apps, on the web and Google Play
Projects
Things I build in my spare time. Most of them run on a server at home, behind nginx, each in its own locked-down container.
Triply
Tinder for travel: swipe through what a city has to offer, then turn your favourites into a day-by-day plan. Places from OpenStreetMap, photos from Wikimedia, AI day plans.
Repeatly
An Android habit tracker with streaks, XP, 100+ challenges, and head-to-head battles with friends over a QR handshake. Offline, no accounts, no subscriptions.
The Ramen Index
Where the ramen is, the cool kids follow. A 3D globe of 17,674 ramen shops in 106 countries, straight from OpenStreetMap, with rankings per country, city and neighbourhood, and a ramen score for any street.
Planly
Self-hosted project management with a canvas, Kanban, Gantt and a chat per task, in one Docker command.
Chatly
A Discord-style community server with text, voice, video and screen sharing. One Rust binary, one SQLite file, no third parties.
Fragly
A 3D arena shooter in the browser: thirty years of favourite FPS moments, Quake-style movement, one Node process, no dependencies.
Movy
A private alternative to sending large files through Messenger or Gmail: resumable uploads and expiring, QR-coded links.
Dataly
Monitoring for the home server: a security report per site, which ports the internet can reach, and one profile per visitor across every log.
sharepointr
Upload, download and manage SharePoint files from R, like a local folder. Open source, with 12 stars on GitHub.
emlTextExtraction
Pull the text out of e-mails and every attachment inside them, recursively: PDFs, Office files, images through OCR, and e-mails within e-mails.
Toolbox
What I reach for, from the first statistical test to the container it ends up running in.
Machine learning
Patterns & anomalies
Statistical testing
Deep learning & AI
Cloud
Infrastructure
Data
Languages & libraries
Work
-
2023 — now
Chief Data Scientist Danske Bank
Copenhagen
-
2017 — 2023
Chief Data Scientist & IT Architect Koncern IT
- Technical lead and full-stack data scientist: from raw data to models running in production.
- Text classification with NLP and language models; predictive models with supervised learning; pattern analysis with unsupervised learning.
- Ran the team’s Kubernetes platform and kept it secure and compliant; parallelised algorithms.
-
2013 — 2017
PhD Fellow Aarhus University, Health
Aarhus
- New methods for pattern discovery in massively parallel sequencing data, and an elaborate variant annotation tool.
- Taught medical students genetics and PhD students exome sequencing.
-
2013
Associate Quality Engineer Biogen Idec
Hillerød
- Coordinated production change-overs, ensured compliance in large-scale manufacturing, and modernised SOPs with visual work instructions.
-
2009 — 2010
Research Assistant Molecular Biology and Genetics, Aarhus University
Aarhus
Education
Degrees
- PhD, BioinformaticsAarhus University · 2013–2017
- MSc, Molecular BiologyAarhus University · 2009–2011
- BSc, Molecular BiologyAarhus University · 2005–2009
Languages
- Danishnative
- Englishprofessional
- Germanfluent
- Japanesebeginner