Senior Knowledge Graph engineer- data sceitist

12/2023 – offen

Developing graph-oriented retrieval systems and optimizing multi-agent in- teractions to drive advancements in linked data and Knowledge Graph (KG). Leveraging techniques from graph data science and Large Language Models (LLMs) to drive advancements in this domain.
Technologies include:
• LlamaIndex, LangChain, NetworkX, streamlit, RDFLib, OpenAI API, Weaviate.
My contributions:
• Benchmarking solutions for starting an in-house infrastructure for local LLMs.
• Setting up the workflow for domain-specific question-answering with LLMs-KG.

Data Science, Large Language Models, Python

11/2019 – 12/2022

Machine learning implementation for computer-aided engineering (CAE) data, GitHub.
• Knowledgegraph:IntroducingsemanticstosummarizeCAEdevelopmentprocesses
and generalize engineering problems.
– Ontologyandfeatureengineering:Abstractingengineeringproblemsintographs, introducing semantics into the domain (OWL), and building graph databases (Neo4j).
– Graph data sceience: Similarity prediction, community detection, classification, and link prediction with GNN, SimRank, PageRank, Graph2Vec.
– NLP: Keywords extraction, document embedding, and classification.
• GenralML:SupportingCAEworkflowsandreusingexistinganalysisbyapplyingvari-
ous ML methods.
– CNN,transferlearning:Vehiclecomponentdetectiontolabelvehiclecrashim- ages.
– Physically-informed learning: Transfer of design experience between related vehicle development projects.
– Sensor data: Similarity prediction and feature extraction.
• LeadingCaeWebVisdevelopment:CAEweb-basedreportingplatformforadvanced
exploration of results, CAEWebVis. Technologies include:
• Django, REST API, React, Neo4j, Docker, NetworkX, OWL, Protégé, SPARQL, Deep- SNAP, TensorBoard, PyG, Shell script, StanfordNLP.
My contributions:
• Initiated the web development working group and shaped the learning of the team for the new technology.
• Made the first available graph modeling for CAE vehicle safety development process.
• Active collaboration with industrial partners and shaped several proposals and working groups.

Data Science, Engineering data management (EDM), Python

6/2014 – 10/2019

Responsible for infrastructure architecture concerning CAE digitalization and web- based reporting. FE crash analysis, method development, and optimization.
• Launching and maintaining the in-house CAE web-based reporting, 16 people user full-stack developer.
• CAE data modelling in Neo4j, the initial state.
• Web application transfer to Django and Neomodel.
Technologies include:
• Python,XML,XSLT,Javascript,jQuery,Django(backendandfrontend),Neomodel,Git, Shell script, LS-DYNA, Heeds, LS-OPT, Ansa, Meta.
My contributions:
• Named as the inventor of European patent application "Device for suspension of a lamp in a vehicle".
• Improved reporting workflows from static reporting to semantic-based reporting.
• Developed geometric optimization methods and integrated them into the team work- flows.

Python, Django, JavaScript-Frameworks