About
I’m Lukas, a researcher and software engineer from Switzerland. I like building things end to end, from FastAPI and Nuxt applications to the self-hosted infrastructure and language models behind them. Right now I’m learning Rust.
I learn something new every day, and I’m always glad to hear how other people see things. I love helping people with IT, because it gives you so much control over your own data and tools; I write down what I learn in case it helps someone else.
Away from the screen I’m learning archery and the piano, and I read about philosophy and personal development. In another life I was a five-time Swiss champion in Rocket League.
Timeline
-
2024 – today
Research assistant, IWI FHNW
Competence centres AI & Optimisation and Digital Business. Deep reinforcement learning and process integration for hospital patient admission.
-
2024 – 2025
MSc Computer Science, University of Camerino
Double degree with FHNW.
-
2023 – 2025
MSc Business Information Systems, FHNW
Thesis: Concept-Based Human-Machine Knowledge System.
-
2023
Developer for a database benchmarking paper, IWI FHNW
-
2020 – 2025
Swiss civilian service, 11 months
IT in radiology at Kantonsspital Aarau, first-level IT support at UPK Basel, patient administration at Reha Rheinfelden.
-
2020 – 2023
BSc Business Information Technology, FHNW
-
2019 – 2020
Full-stack web developer, SOWACOM GmbH
A web interface for the ERP storage system, after an internship building an internal web shop.
-
2017 – 2024
Rocket League for Servette Geneva Esports and Edelweiss Esports
Player, captain, then manager of the section; the team later became Edelweiss Esports.
-
2015 – 2018
IT middle school, Basel
Publications
- Deep Reinforcement Learning in Patient Admission Process to Optimize Bed Occupancy and Revenue. L. Jakober, T. Hanne, R. Dornberger. IEEE ISCBI 2025, conference paper
- Optimizing the Integration of the Patient Admission Process in a Swiss Rehabilitation Center with Multiple Parties. L. Jakober, D. Wörner, T. Hanne. IFAC-PapersOnLine 2025, conference paper
- Concept-Based Human-Machine Knowledge System. L. Jakober, P. Christen. IEEE SMC 2025, poster
Skills
- Every day
- Python with FastAPI · PostgreSQL · Nuxt (Vue.js)
- Machine learning
- PyTorch · scikit-learn · pandas · NumPy · spaCy · Stable Baselines3
- Infrastructure
- Linux (Fedora / RHEL) · Podman · Ansible · Keycloak · OpenBao · OpenStack with OpenTofu
- AI / LLMs
- Self-hosted inference with vLLM and llama.cpp · coding agents
- Business
- Project and process management · business intelligence · business-IT alignment
- Languages
- German (native) · English (B2 First) · French (DELF B1) · some Mandarin
Contact
Always happy to hear about interesting projects and other perspectives.