• EOA team at Agroscope (since 01/2023)
• Support and maintenance

• Zum Löwenzorn AG (since 03/2020)
• Support and maintenance (Nextcloud)

• Agroscope (since 05/2018)
• Support and maintenance

• EOA team at Agroscope
• GPU accelerated Data Science infrastructure

• Agroscope
• Migration of Data Science infrastructure

• Crowd Container
• Data exploration & visualisation

• Agroscope
• Git training

• Confidential
• Application containerisation

• wisy (until 02/2020)
• Data Science infrastructure (Operation and maintenance)

• Innosense
• Testing of Gridsense-algorithm

• Agroscope
• Data Science infrastructure

• Zürcher Nachhilfe (until 12/2019)
• Prediction model
• Database design

• Innosense (until 09/2018)
• Data Science infrastructure
• Support and maintenance

• Parfumaktion
• Data exploration & visualisation
• Customer segmentation

• Alpiq InTec
• RStudio Server training
• Templates for Data Analysis

• Confidential (until 03/2018)
• Database design
• Matching algorithm

• zkipster (until 01/2019)
• Data Science infrastructure
• Social network analysis

• PostMail
• Process optimisation (manual letter sorting)

• Multi-arch¹ dev container
• Mojo dev container

• Multi-arch¹ docker images
• JupyterLab MAX/Mojo docker stack
• MAX/Mojo docker stack

• GPU accelerated multi-arch¹ dev containers
• CUDA-based Data Science dev containers

• GPU accelerated multi-arch¹ docker images
• CUDA-based JupyterLab R docker stack
• qgisprocess
• CUDA-based R docker stack
• qgisprocess
• CUDA-based JupyterLab QGIS docker stack
• CUDA-based JupyterLab Python docker stack
• CUDA-based Python docker stack
• CUDA-based JupyterLab Julia docker stack
• CUDA-based Julia docker stack

• Multi-arch¹ docker images
• JupyterLab QGIS docker stack

• GPU accelerated multi-arch¹ docker images
• CUDA-based JupyterLab R docker stack
• CUDA-based R docker stack

• Multi-arch¹ docker images
• JupyterLab Julia docker stack
• pubtools
• JupyterLab Python docker stack
• R docker stack
• base
• Julia docker stack
• base, pubtools
• Python docker stack

• Containerised installations
• Git LFS

• Containerised source-installations
• R
• Git

• Multi-arch¹ docker images
• Docker buildx qemu (archived)
• JupyterHub onbuild
• JupyterHub
• GHC musl
→ Used to build the statically linked Linux amd64 and arm64 binary releases of pandoc.

• Deployment templates
• IDE for Data Scientists: Jupyter
• JupyterLab + code-server

• Multi-arch¹ docker images
• Docker buildx (archived)
• JupyterLab Julia docker stack
• JupyterLab R docker stack
• Julia docker stack
• R docker stack

• Docker images (linux/amd64)
• Version-stable Rocker images (archived)

• Templates for R
• API: plumber
• Best practice
• packrat
• pure R

• Deployment templates
• IDE for Data Scientists: RCloud (archived)
• DevOps Platform: GitLab CE
• Reverse proxy: Træfik

• Docker images (linux/amd64)
• RCloud (archived)
• RCloud search (archived)
• RCloud SessionKeyServer (archived)

Development of prediction algorithm
A core feature of the wisy app is the prediction algorithm for calculating a students passing probability for an exam.
Olivier has taken on the task of creating a solid prediction model with R. And the provision as a REST API makes the implementation a breeze.

IT Operations
Operation and maintenance of wisy's GitLab server to manage the entire DevOps lifecycle with one application. Automatic build and test of docker images for the wisy app and prediction API.

Operation and maintenance of wisy's application servers and automatic deployment and scaling of docker services in Swarm mode.

IT tool set: R, SQL, Nginx, Apache, MySQL, Git, GitLab, Docker

2010 – 2015: Implementation of various People Counting Systems
Olivier was responsible for the overall project management of more than a dozen People Counting System installations.
He ensured that the various components used (Harware, Software) and interfaces to third-party systems were correctly implemented.

IT tool set: Various video-based People Counting Systems, IIS, PostgreSQL, Eclipse, SVN, Mantis Bug Tracker, Windows

2009 – 2010: Optimisation of time needed for leaving a car park near a stadium
In order to avoid long waiting times in front of the cash registers or in the car park after events, the operator had an information system installed, which informs the visitors about the expected time needed for leaving the car park.

Olivier lead the entire development process of the product. It measures the flow of visitors, calculates them expected waiting time, displays it on a website and sends a signal to the traffic control system.

IT tool set: Apache, PostgreSQL, MySQL, Eclipse, SVN, Debian

2008 – 2009: PostMail REMA (REengineering MAilprocessing)
Support of the application manager in the requirements engineering of the rules of the Production and Planning System (PPS). Olivier took on the interface function between IT Post and the supplier. He specified the rules of individual PPS modules as well as the integration, regression and acceptance tests.

Miscellaneous
Full responsibility for the internal IT infrastructure at ASE GmbH. This also included the operation and maintenance of the enterprise and application servers.

IT tool set: Apache, IIS, PostgreSQL, MySQL, MS SQL, VirtualBox, Debian, Windows, TYPO3, Mediawiki

Setup of extensive simulation models
With its simulation models, Swiss Post was able to demonstrate the dependency between the offered services, letter volumes to be processed, deployed resources (personnel, machines) and throughput times.
Team Simulation had the task of discovering bottlenecks and weaknesses of the design and to work out marketable solutions.

Simulation of commissioning and letter container storage
Before these fully automated areas of the letter sorting centre were built, Swiss Post agreed with the suppliers on preliminary simulation studies.
Team Simulation was responsible for generating the load data sets, performing the simulation studies and optimizing the steering algorithms.

Defining the ruleset of the Production and Planning System (PPS)
Olivier was a member of the core team, which was responsible for the requirements engineering of the PPS rules. His task was to translate the steering algorithm used in the simulation model into requirements for PPS.

Testing the modules Monitoring and Controlling of PPS
Based on the ruleset, Olivier specified the regression tests of the modules Monitoring and Control. As part of the test automation, he created a load data generator for the supplier.

IT tool set: R, XML, UML, Eclipse, SIMPRO, Mantis Bug Tracker

On site for Swisscom and other leading telecom providers
Installation of transmission systems, predominantly in the SDH transmission technology. Setup, cabling (LF, HF, coax and fibre optic) and connection of network elements such as add/drop multiplexer and ADSL equipment.

On site for TV studio Zurich
Installation of television equipment, cabling and wiring.

Major subjects
Computer-aided Engineering (CAE)
Process Design

Accompanying «Berufsmittelschule»