Text Mining Experts in Germany

• Building, implementing and selling automated Data Science solutions such as Scorecard Factory and Forecast Factory
• Implementation of automated end-to-end cloud processes
• Development of LLM and NLP models
• Creation of interactive reports
• Support for national and international large corporations as well as medium-sized companies in implementing ML projects

• Development High-Performance ad distribution via auction
• Holistic (multi-campaign & multi-channel) advertisement placement optimization
• Algorithmic optimization for NP-Hard/NP-e
• Multiple Knapsack Problem with constraints
• Online estimation of parameters in stochastic environments

Tools: Python, R, Kotlin, MILP/SAT/CP Solvers, Pytorch, Pandas, Docker

• Calculation of attribute importance based on driver analyses
• Interpretation, reporting, and consulting

End-to-end responsibility for all products – conception, architecture, development, and production operations as the sole developer; also customer discussions, proposals, and marketing.

Underwriting Copilot – AI assistant for industrial insurance (in production with a customer)

• Supports underwriters in analyzing industrial insurance submissions – in production at an industrial insurer.
• Framework-independent RAG architecture with Hybrid Search (BM25 + pgvector) across large, heterogeneous document collections.
• Two-stage evaluation and observability pipeline (code assertions + LLM-as-Judge) that makes answer quality, retrieval accuracy, and citation integrity measurable in a regression-safe way.

Kaiki Menu Analyzer – Data intelligence platform (in production with a customer)

• Automatically collects and analyzes food data from around 25,000 German restaurants.
• Scalable 7-container architecture (FastAPI, partitioned PostgreSQL, Redis/RQ) with LLM-supported extraction of structured data from PDF, HTML, and images.
• Complete CI/CD pipelines (GitHub Actions), production cloud deployment, interactive dashboards (Dash).

Kaiki GEO Atlas – GEO platform (in production with a customer)

• Measures brand visibility across five AI engines (ChatGPT, Gemini, Perplexity, Grok, Claude), each augmented with web search, orchestrated as a DAG workflow pipeline (Dispatcher → Sub-Workflows → Scoring → Report) with fail isolation.
• 6-container deployment (FastAPI, Celery, Redis, PostgreSQL); LLM cost estimation, PDF audit report, rule-based cross-signal insights (no additional LLM effort).

Data Pipeline & Analytics Platform – Competitive analysis in the automotive aftermarket

• Automated data pipeline with gap analysis algorithms and role-based access control; 230+ tests.
• Backend with FastAPI, PostgreSQL, SQLAlchemy.

Product development (actively in development)

BankingGPT – AI assistant for complaint management in cooperative banking

• Security architecture at its core: no AI draft reaches the customer without human approval – the approval decision is stored in auditable code, not in the language model (monotonic: the model may increase the level, but never lower it).
• Genuine agentic components, each with its own boundaries: the model selects tools itself through an MCP server (read-only, allowlisted, capped, fail-safe); sensitive cases are sent via the open A2A protocol (JSON-RPC, Agent Card, message/send/tasks/get; client implemented independently) to a standalone specialist agent (securities/law), which never lowers the review requirement (pinned by tests).
• Evaluation-driven across ten analysis rounds; an independent review and blind tests uncovered a security defect that nine automated runs had missed.
• Voice AI frontend, responding live: covered cases are answered during the conversation, while sensitive cases are escalated before generation; measured response latency < 7 s (local GPU STT/TTS).

Stack & production readiness: Python, pydantic-ai, FastAPI/Celery, PostgreSQL/pgvector, FastMCP, fasta2a, Docker; multi-tenant (physical vector isolation per tenant), encrypted PII, OWASP-LLM-reviewed, 275 tests, CI/CD; provider-portable (Ollama / EU-Cloud-Vertex).

After-Sales Assistant – agentic RAG/GraphRAG assistant based on public OEM manuals (automotive after-sales)

• Genuinely agentic with LangGraph: ReAct agent with four tools and conversation memory – the model decides itself whether to use the manual (RAG, Chroma), a knowledge graph (GraphRAG, Neo4j/Cypher – decoding warning lights), or a workshop/booking service.
• Human-in-the-loop before the irreversible action: before every appointment booking, the graph pauses (interrupt) and obtains the driver's explicit confirmation – the same approval-before-action discipline as in BankingGPT, in a different framework.
• Eval as a CI gate: three-part scorecard (RAGAS grounding + deterministic tool-routing accuracy + DeepEval safety: does the answer mention the warning first in the case of a critical warning?) blocks the pipeline; provider-agnostic (OpenAI/Azure/Anthropic), FastAPI with token streaming.

Stack: Python, LangChain/LangGraph, Chroma, Neo4j, RAGAS/DeepEval, FastAPI, Docker.

• Assist with the development of system requirements and specifications to ensure requirements are testable and 21 CFR Part 11 requirements are met
• Coach implementation teams in the proper execution of validation documents
• Evaluate proposed changes to validated computer systems and recommend level of validation activities required
• Coordinate audits of internal computer systems validation activities, protocols and procedures, and prepare responses
• Identify and qualify all computer systems impacting cGMP operations using a risk-based methodology
• Develop CFR Part 11 computer systems validation plans, qualification test protocols, traceability matrices, reports, IQ/OQ protocols and all deliverables within the scope of the validation plan
• Develop and maintain test plans, test scripts and user acceptance tests and manage their execution
• Act as CSV lead for all validation projects and execute or oversee validation plans and documents
• Perform project management activities for the CSV process within the scope of system projects
• Work with project manager to include validation activities in implementation timelines
• Manage internal CSV resources to facilitate completion of qualification activities
• Ensure initiation, preparation and closeout of all CSV-related deviations, discrepancies and change control documents
• Work closely with Validation Manager and QA Compliance to ensure appropriate validation of cGMP computer systems
• Conduct or facilitate validation and 21 CFR Part 11 training

• Designed a future-proof client database for marketing purposes
• Analyzed requirements, designed, and modeled an entity-relationship model
• Consolidated and optimized a client file from various data sources for targeted marketing campaigns
• Worked closely with marketing and IT in an agile environment to iteratively develop the solution
• Technologies and methods: MS Office (mainly Excel), MS Dynamics CRM, MS SharePoint

• Developed ASR and TSR-based speech processing pipelines on AWS, enabling efficient feature extraction and scalable deployment for speech and text analytics.
• Built a Multimodal Speech Emotion Recognition system combining NLP and deep learning (audio + text), achieving 98% accuracy and supporting real-time, cloud-based inference.
• Designed and optimized end-to-end model training and evaluation workflows using AWS services (S3, EC2, Lambda) to ensure performance, reliability, and reproducibility.
• Created and deployed interactive, user-friendly dashboards for data visualization and insight generation, supporting research teams and management in data-driven decision-making.

• Conducting research at the interface of proteomics and artificial intelligence, focusing on the application of machine learning models (e.g., neural networks, clustering algorithms, and feature extraction) to analyze complex biological datasets.
• Developing and teaching AI-based analytical workflows for molecular and proteomic data, integrating tools such as Python (scikit-learn, TensorFlow, Pandas) for predictive modeling and data visualization.
• Collaborating with interdisciplinary teams to explore data-driven hypotheses in molecular genetics and enhance biological interpretation through AI-assisted pattern recognition.
• Implementing automated data processing pipelines to improve reproducibility and FAIR data management in high-throughput experiments.

• Lead engineer for the MatchMaker Toolbox (KNIME): Index Builder, Approximate Matcher, Character Mapper, license nodes
• Designed M|ARS (MatchMaker Agentic Retrieval System) — hybrid retrieval combining deterministic search + LLM tooling
• Developed internal RAG and search prototypes (MatchMaker + vector search + LLM)