Freelancer in Data Science

Within this time I was taking my budget for a long-term traveling experience and started my new path as freelancer. I built a van, travelled south America and worked in remote projects as well as started my own project.

I continuously educated myself in data science and software development, learned new technologies and worked for customers in the area of machine learning, optimization and LLM training.

Successes:

• Successful implementation of cloud-based data crawler
• Data-base migration with 85% cost reduction
• Solar system optimization with up to 35% cost reduction
• RFID tag recognation modeling, model prediction quality increase of 15%

Function:

• LLM training and validation data creation
• Home solar system cost optimization
• Machine learning for RFID tag classification
• Data crawling and evaluation
• Van construction and crafting
• Traveling guide

Tools and technology: Python (sklearn, pandas, numpy, Django, sqlAlchemy, pyTorch), Docker, JavaScript, MS Azure, AWS EC2, Lambda, S3, MySQL, Hadoop & Spark, LLM Prompting

After successfully rolling out the software, I switched to combustion engine development. We developed and applied algorithms for engine combustion control and control unit software development for Japanese customers. Additionally we developed processes and methodologies to strongly reduce the engine development time and costs for our customers.

Success:

• Modelling of prototype engines, strong reduction of engine calibration effort

Function:

• Mathematical methodology development
• Machine Learning and Optimization expert
• Technical expert for methodology and combustion control

Tools and Technologies: Matlab/Simulink, Python (sklearn, pandas, numpy, pyTorch), CI/CD, Jenkins, Git, Jira, Confluence, ETAS INCA, AVL PUMA, A&D ORION

Next to my PhD thesis research, I worked at IAV GmbH in the department for method development for combustion engine calibration. I was in charge to develop algorithms for an adaptive test design method for the optimization of engine tests to train machine learning models with the target to optimize the combustion.

Additionally, I was the technical expert to adapt a commercial software, which was equiped with my calculation logic and developed ML algorithms.

Successes:

• Commercial software release with ML algorithms
• Engine calibration time reduction up to 30%

Function:

• Mathematical method development (linear algebra and stochastic)
• Expert for test design algorithms
• Complex calculation logic development
• Consulting expert for the software development team

Tools and Technologies: Matlab, JAVA, SQL, CI/CD, Jenkins, UML, Socket Interface, Jira, Git, StarUML