Converted a monolithic software in the pharmaceutical sector from an on-premises system to a web application.
Tools: Python, VueJs, Quasar, Docker
Oct 2022 - Present
2 years 10 months
AI Platform
R+V Versicherung AG
Built an AI platform for semi-automated processing of large data volumes with machine learning and ETL pipelines in a large team of developers and data scientists.
Secure handling and explainable black-box processing of sensitive data.
Monitored systems and visualized processes using monitoring and reporting tools (ELK).
I advise clients on data science and machine learning projects.
For the largest client, R+V, I design in an eight-person team the architecture for ETL pipelines of an AI platform for semi-automated processing of sensitive customer data.
Modernize the project's tech stack with Docker and Elasticsearch and train my colleagues in monitoring and reporting with the ELK Stack.
For another client that builds golf courses, I develop an ML application for quick detection of lawn diseases (using a fine-tuned MobileNetV2 and semantic segmentation).
In a completed project, I led the team and managed the project to convert an existing monolithic on-premises system into a web application.
Aug 2021 - Jul 2022
1 year
Site Search
RWE AG
Consulting to optimize the search functionality of the corporate website and related data engineering processes.
Improved user experience with search suggestions and input completion.
I worked on both in-house software products and client projects.
Played a key role in developing the company's main product, a search technology ("Rocket Search", 05/2019 - 10/2022): The software replaces the search bar on the client's website, crawls page content and indexes it in Elasticsearch to present it to end users in an intelligent search function. Keywords include result quality, improved search suggestions ("did you mean?"), autocomplete, filters, and a portal where user behavior is tracked and analyzed in a dashboard. You can also make settings such as boosting certain keywords, creating synonyms, or setting exclusions. Tools: Elasticsearch, Kibana, Scrapy, Python, VueJS, Flask, Docker, nginx, Gitlab Runner.
Developed the internal "Sales Tool" (07/2021 - 03/2022): A single page application for tracking and analyzing internal sales processes and customer management. Tools: Elasticsearch, Kibana, VueJs, Python, Flask, Docker, nginx.
Completed an open source microservice software ("Roxcomposer", 10/2018 - 04/2019): Developed a framework for creating machine learning pipelines that outsources steps like tokenization into reusable microservices and lets you assemble them via a GUI. Tools: Elasticsearch, Python, Django, scikit-learn, nltk, gensim.
In client projects I also worked daily with the ELK Stack, developed ETL pipelines, and built full-stack web applications for data processing and visualization.
Jan 2016 - Dec 2019
4 years
Mainz, Germany
Student Research Assistant
Johannes Gutenberg-Universität
Tutor for the courses Artificial Intelligence, Programming Languages, and Software Development.
Jan 2016 - Dec 2016
1 year
Mainz, Germany
Student Research Assistant
Universitätsmedizin
Analysis of patient data (neurological DTI scans) from patients with multiple sclerosis.