AI & Knowledge Discovery Leader · Leipzig, Germany

Bjoern Oest Hansen

I build scientific AI, knowledge graphs and discovery platforms that turn fragmented research knowledge into better decisions.

Currently leading the Knowledge Discovery & AI team at KWS, and helping train new scientists through its Data Science Academy.

Get in touch LinkedIn Google Scholar
A knowledge graph linking literature, genes, traits, experiments, language models and biological networks Literature Genes Traits Experiments Language models Networks Knowledge graph
Fig. 1: From literature to traits, connecting evidence to find the genes that matter.

01 Impact

What I've built

  • 10+years building scientific AI and discovery platforms
  • 5people in the team I lead
  • 10AI capabilities delivered: 4 AI solutions and 6 ML methods
  • 200+species integrated in a cross-species knowledge graph

  • A Knowledge Discovery & AI function

    Set up KWS's Knowledge Discovery & AI function, including its strategy, roadmap and ways of working.

  • A portfolio of scientific AI

    Delivered AI solutions and machine-learning methods that support biological target discovery and prioritisation.

  • A discovery platform

    Combined knowledge graphs, scientific literature, biological data, machine learning and AI-assisted hypothesis generation in one platform.

  • Adoption across R&D

    Worked with scientists, breeders, data science, digital and IT to make AI workflows practical in day-to-day research.

02 Leadership

How I lead

  • Team

    Building and leading a team

    I lead a team of five data scientists and bioinformaticians, four of whom I hired. I have also led several R&D projects.

  • People

    Developing new scientists

    I lead curriculum design and mentoring for KWS's Data Science Academy, an internal graduate programme for new scientists (three participants so far). I have also hosted master's students and interns.

  • Strategy

    Aligning work with company strategy

    I developed the strategy for in-silico gene discovery and knowledge discovery in R&D, so that our team's work and priorities line up with KWS's overall strategy.

  • Governance

    Responsible use of AI

    I led R&D's input from the technical perspective to KWS's AI compliance guidelines.

03 Background

From the bench to building AI functions

I started as a bench scientist studying gene co-expression networks before moving into computational biology, then founding my own knowledge graph startup, OmicsDriven, and eventually setting up and leading KWS's Knowledge Discovery & AI function.

That path, from wet-lab research to data science to organisational leadership, shapes how I work today: I still get hands-on with knowledge graphs, machine learning and generative AI, but spend most of my time turning those capabilities into platforms, workflows and team structures that scientists and breeders actually trust and use.

04 Experience

Career & education

KWS SAAT SE & Co. KGaA

2018 – present

  1. Research Lead → Senior Research Lead, Knowledge Discovery & AI

    2023 – present

    Leading the Knowledge Discovery & AI function, applying scientific AI, knowledge integration and analytics to candidate-gene discovery, hypothesis generation and research decision-making.

    • Lead a team of five data scientists and bioinformaticians, four of whom I hired.
    • Developed the R&D strategy for in-silico gene discovery and knowledge discovery, aligned with KWS's overall strategy, and set delivery roadmaps with research, breeding, digital and IT.
    • Oversaw discovery workflows combining literature-derived evidence, knowledge graphs, TWAS, biological networks, experimental data and ML predictions.
    • Introduced LLM and RAG capabilities into research workflows, including AWS Bedrock solutions integrated with Microsoft Teams.
    • Led R&D's technical input to KWS's AI compliance guidelines.
    • Lead curriculum design and mentoring for the Data Science Academy, KWS's internal graduate programme.

  2. Data Scientist → Senior Data Scientist, Advanced Analytics

    2018 – 2023

    Computational approaches for candidate-gene discovery and biological target prioritisation.

    • Developed machine-learning and data-integration approaches for in-silico candidate-gene identification.
    • Built gene-trait extraction systems, biological knowledge graph capabilities and graph-based prediction methods.
    • Integrated TWAS, biological networks, literature-derived knowledge and experimental evidence into target prioritisation.

OmicsDriven

2017 – 2021

  1. Founder

    Built a knowledge graph platform in Neo4j integrating biological networks and scientific knowledge across more than 200 species, for knowledge exploration and gene-function prediction. Owned product design, technical development and user engagement.

Earlier roles

  • Scientific Employee, Medical Informatics · Medical Research Centre Göttingen
    2017 – 2018
  • Scientific Employee · Humboldt University Berlin / TOPOI
    2016 – 2017

Education

  • PhD research · Max Planck Institute of Molecular Plant Physiology
    Integrative transcriptomic analysis of co-expression networks (not submitted)
  • MSc & BSc, Biology–Biotechnology · University of Copenhagen

05 Expertise

Skills & tools

Leadership

  • Team leadership
  • Hiring
  • Mentoring
  • Curriculum design
  • R&D strategy
  • AI governance
  • Project leadership
  • Stakeholder management

Scientific AI

  • Machine learning
  • Generative AI
  • Large language models
  • NLP
  • RAG
  • Recommendation systems

Knowledge & discovery

  • Knowledge graphs
  • Information extraction
  • Semantic search
  • Evidence integration
  • Hypothesis generation
  • Target prioritisation

Computational biology

  • Candidate gene discovery
  • Gene function prediction
  • Biological networks
  • Multi-omics integration
  • TWAS

Platforms & engineering

  • AWS Bedrock
  • AWS
  • Azure
  • Neo4j
  • MLOps
  • Data engineering
  • Python
  • SQL

Languages: Danish (native), English (fluent), German (intermediate), Spanish (basic)

06 Publications

Publications & talks

Full publication list on Google Scholar

07 Contact

Let's talk.

Scientific AI, knowledge discovery or building an R&D data capability: I'm always happy to hear from people working on these problems.

bjoernoesth@gmail.com