Supporting the team with the ever-growing demand on devops topics in more and more security-aware deployment environments.Since November 2023 also following the very exiting and thriving path of AI and data analysis within the Airport business:- successful prediction analysis about VTT (variable taxi time) and Passenger forecasts with Python ML stack, partially on data with >10 million records. This includes XGBoost with hyper parameter tuning, scikit-learn, feature engineering challenges including label encodings (e.g. for taxi stands, runways, callsigns, dates), feature importance evaluations, matplotlib.- evaluation of detection rates for NER (named entity recognition) in the context of airport service confirmations (document detection and recognition). Just for fun evaluated reliability for tail number recognitions. - comparison of various error metrics for legacy passenger forecast calculations in SKYport with ML-based approach. - first experience with MLOps (MLFlow, ONNX model creation and runtime use in existing J2EE environment) - first expertise with LLMs and RAG (Hugging face models, TEI (Text Embedding Inference), CUDA Python environment, Ollama) based on division knowledge (JIRA,Confluence, GIT commit history, PDF/Word documents) - current evaluation of manual (and fragile) prompt engineering techniques versus DSPy to build RAG-based 'knowledge chatbot' and intended legacy code transpilations. - Use of OLAP ClickHouse database (using its vector search and inverse index full-text search functionality together with SQL-based data fusion) for RAG.