NLP/LLM Chatbot

• Conceptualized and implemented an LLM-based case assistant (file assistant).
• Selected and evaluated RAG methods; designed hybrid RAG information retrieval using Elasticsearch + embeddings.
• Built ingestion pipelines for multiple document formats; analyzed and aligned with source systems.
• Developed a Streamlit-based chatbot GUI and performed NLP-based causal chain analysis for regress cases.
• Evaluated analytical LLM methods; deployed via Jenkins to OpenStage.
• Technologies used: Streamlit, Elasticsearch/OpenSearch, Hugging Face, Python, OpenAI API, Jenkins

• Worked in a security-cleared (SÜ2) and BSI-compliant environment.
• Explored and advised on Knowledge Graph use; designed & implemented pipelines with Neo4j.
• Raw data extraction from SQL dumps and mapping to entities; graph data model design.
• Advanced NLP: NER, topic & event extraction, language detection; extraction of complex relations using POLE (Persons, Objects, Locations, Events).
• Applied graph algorithms (incoming degree, community detection – Louvain/triangle count) to identify relevant entities and relationships.
• Integrated GeoIP/domain extraction and URL evaluation; implemented streaming ingestion (Kafka).
• Built an LLM/Chat Streamlit app deployed on Cloud Foundry on Kubernetes; integrated into BSI-compliant environment.
• Technologies used: Neo4j, Pandas, Polars, spaCy, OpenNLP, Hugging Face, Kafka, Cloud Foundry, Kubernetes, GitLab CI/CD, Streamlit

• Led development of large-scale AWS-based data pipelines for multiple datasets (Glue/PySpark, Athena, S3, Lambda, IAM, SQS/SNS, DynamoDB/DAX, Elasticsearch).
• Processed up to 30+ billion rows; implemented automated restart & recovery procedures and unit tests.
• Architected daily/historic/housekeeping pipelines; performed production problem analysis and internal advisory.
• Performed cost analysis, derived measures, implemented cost savings, and monitored reductions.
• Refactored complex vehicle master data (7 datasets, ~35 tables; mid-level business logic; >1B rows/day).
• Ran feasibility analyses, requirements refinement, and implementation concepts; authored documentation.
• Technologies used: AWS Glue, PySpark, Athena, S3, CloudWatch, Lambda, SSM, KMS, IAM, SQS/SNS, Elasticsearch, DynamoDB/DAX, EC2 (tests), Terraform

• Designed & implemented data model for passenger revenue (ticket sales and other services) considering AWS specifics (sort key, data distribution).
• Built multi-stage Data Warehouse architecture: two-stage staging, Core (persistent business model), Analytics (optimized for SAP BO), and Output for file interfaces.
• Designed ETL/Data Pipelines: daily, historic reprocessing, housekeeping; Lambda/S3 events for job control; reload from raw data; processed ~1M rows/day.
• Implemented jobs with AWS Batch (Python, Docker) using SSM Parameter Store; integrated DynamoDB; limited Red Hat OpenShift on AWS.
• Managed build pipelines via GitLab CI/CD for >15 projects (dynamic parameterization, Docker images, build scripts, Python merge/deploy).
• Conducted performance analysis and workshops with AWS Support; optimized Redshift WLM; ensured operations stability with weekly coordination & troubleshooting.
• Technologies used: AWS Redshift, Lambda, S3, Batch, Docker, DynamoDB, Python, Pandas, SAP BO, GitLab CI/CD, Flyway

• Analyzed materials control and logistics planning; prioritized with stakeholders; investigated the SAP–DWH interface.
• Investigated Teradata SQL, ETL modules, views, and data models across source, inbound (SAP→DWH), and DWH layers; proposed solutions and documented results.
• Technologies used: Teradata SQL, ETL analysis, SAP DWH

• Prototyped collaborative VR space with messaging, user management, and real-time constraints; explored scaling and availability; visual concept and 3D modeling.
• Technologies used: Akka, Scala, C#, Unity (VR GUI), Blender, WebRTC

• Designed and built a hybrid AI storytelling/world-building platform combining knowledge graphs, LLMs, and custom entropy-shaping heuristics.
• Implements multi-layer graph spaces and Graph RAG; integrates generative services for text, image, and music; React-based graph explorer.
• Security & deployment with AWS AppSync/Amplify back ends, REST/GraphQL endpoints, CI/CD, AWS CDK, Cognito/IAM.
• Technologies used: Neo4j, Qdrant, Python, React, GraphQL, AWS AppSync/Amplify, AWS CDK

• Analyzed impacts of the One.ERP Integration on DWH-T systems; mapped inbound/outbound data flows across SAP and BW landscapes.
• Assessed core finance processes: revenue processing, month-end close, dunning/receivables, logistics, sub-ledgers, billing/invoicing.
• Prepared knowledge materials on Controlling, Treasury, Accounting, IFRS/HGB differences; clarified KPIs (OCF, ROCE, CapEx, EBITDA).
• Technologies used: SAP BW, SAP ERP, Oracle, DWH-T, ConBW/IKOS

• Consulting across BI/DWH architecture and reporting for banking functions with emphasis on Controlling and risk-related reporting.
• Technologies used: SAP BW, Oracle, QlikView

• Co-founded and managed agency operations; production planning; delivered multiple client projects; software & multimedia engineering.

Abitur (German A-levels), Werner Heisenberg School, Rüsselsheim