Qwen Experts in Germany

• Complete greenfield rebuild across the whole stack — built AI-native: backend in Go and NestJS, PostgreSQL (CNPG) on K3s with GitOps/Terraform; frontend on Angular 22, zoneless.
• Orchestrated coding agents (e.g. Claude Code, Cursor) across the entire lifecycle — architecture, implementation, testing, reviews, documentation — driven by Specification-Driven Development (SDD).
• “Bruno” — LLM commentator persona backed by RAG and MCP for a personality that stays consistent across all generations (match previews, post-match reports, his own virtual bets).

Angular 22 (zoneless, without Zone.js), Claude Code, Claude Code Skills, CNPG, Cursor, Design Tokens (Spec for Code), Docker, Gherkin, Git, GitLab, GitOps, Go, Google Gemini, Grafana, Hetzner Cloud, K3s, Keycloak, Kubernetes, Lighthouse, LLM Integration, Model Context Protocol (MCP), NestJS, Node.js, NPM, Playwright, PostgreSQL, Prometheus, RAG, REST, Specification-Driven Development (SDD), Structured Outputs, Terraform, TypeScript, Vitest

• Build and operate an end-to-end MLOps platform on Azure ML and Kubernetes (Kubeflow) for the automated deployment, monitoring, and scaling of forecasting models (including Temporal Fusion Transformer, Informer, Autoformer).
• Implement CI/CD pipelines in Azure DevOps for the full ML lifecycle – from resource provisioning (Terraform), data transformation (Hugging Face Datasets, Pandas, PyTorch, CUDA cluster) through training and evaluation to model registry and endpoint deployment.
• Integrate MLflow for experiment tracking, model versioning, performance monitoring, and automated registration in the Azure Model Registry.
• Develop and containerize PyTorch training jobs (Azure Notebook, Jupyter Notebooks) for price and time series forecasting (PFC models) with automatic rollout via Azure ML Endpoints and REST/gRPC interfaces, Docker containerization, secured with OAuth 2.0.
• Set up monitoring and alerting mechanisms (Prometheus, MLflow Metrics), log centralization, and cost monitoring.
• Automate infrastructure provisioning and model deployment using Terraform, Helm, and Azure CLI; connect to existing market data systems and event pipelines.
• Migrate existing workloads and databases (IONOS → Azure, MongoDB) with integration into central MLOps workflows and internal networks.
• Extend the platform with LLM-based tools (LangChain, LangServe) to integrate GPT-based analysis modules into existing Spring Boot services for market anomaly detection and automated reports.
• Analyze and architect a software solution to process large volumes of data efficiently (>3000 messages/sec.) (market data store).
• Spring Boot / Java 21 container development with RabbitMQ for distributing stock market data via MongoDB (Kubernetes) with fast storage of data in Redis RMaps, deduplication, forwarding messages to Read Model queues, and building Read Models for UI display in MongoDB.
• Integration of RESTHeart to create a REST API for MongoDB.
• Build an Angular frontend to simplify data queries and master data maintenance.
• Agentic coding with remote and local LLMs (Claude Sonnet, Ollama Qwen) and MCP servers.
• Develop Python scripts for transforming and cleaning incoming stock market data (Pandas, scikit-learn).

• Consulting and project management to build an AI-based automation platform for market data analysis
• Automated testing of financial data
• Market data automation with Python and N8N
• AI data integration and AI learning
• Use of Ollama and Qwen
• Data organization and database management

• Fine-tuned and deployed Generative AI and LLM models (OpenAI, DeepSeek, Qwen-2.5) using PyTorch and Hugging Face, increasing ERP automation accuracy by 25%.
• Designed and implemented a secure RAG-powered AI Chabot for customer-specific invoice and quotation generation, cutting response times by 40%.
• Architected cloud-native AI/ML pipelines on AWS and GCP with Docker and Kubernetes for scalable model training, deployment and monitoring.
• Developed and integrated an API-driven AI Chabot (Telegram) with ERP systems, boosting document processing speed by 30%.
• Built AI agents for chatbots to enable multi-step reasoning, intelligent task execution, and context-aware interactions.
• Applied ML and NLP techniques for intelligent document understanding, workflow automation, and data-driven business decisions.

Freelance work for a large German family office

Built and further developed a modular B2B SaaS platform for professional wealth management and family offices, alongside freelance delivery of production-ready AI, data management, and automation solutions for the wealth management sector.

• Developed an AI governance framework for regulated finance and asset management workflows, aligned with DORA, BaFin-related governance expectations, and data protection requirements, including role definitions, access levels, and decision rules.
• Designed an agentic system with a locally operated open-source language model, including Qwen2.5 via Ollama, for secure querying of an asset database through text-to-SQL-to-text workflows.
• Built AI-supported analysis and reporting capabilities that generate structured answers, tables, and charts from asset data, with domain validation through resolver logic and RAG elements.
• Developed production-grade data import workflows for financial service provider data from CSV, PDF, and API sources, including validation, plausibility checks, and reconciliation with existing asset data.
• Solved the asset matching problem without a cross-system primary key through multi-stage validation rules and human-in-the-loop approvals.

Results:

• Secured EUR 250,000 in SaaS revenue in 2024, exceeding the forecast by 20%.
• Acquired family office clients with EUR 1.6bn in assets under management.
• Reduced manual effort for the largest client by approx. 3 days per month through automated data import and reconciliation processes.
• Reduced operational error risk through structured data validation, multi-stage asset matching, and human-in-the-loop approvals.

• Consultant for system architecture, AI agents & integration, coach for data & process logic, Graph-RAG approaches, security and data protection.
• On-premise AI solutions with high compliance and performance requirements.
• Architecture decisions, operational setup, strategic prioritization & deployment.
• Technologies: LiveKit JS SDK, LiveKit Agents, Web Audio API, JS, AudioWorklet, Loki, vLLM, Zscaler, Docker, Neo4j, MySQL, Python.
• Models: GPT-OSS 20B, Whisper large v3 turbo, Qwen3-TTS.

• Research and develop models for chatbots, NLP and LLMs (e.g. Llama, Qwen, OpenAI)
• Enhance chatbots with RAG, in-context learning
• Supervised fine-tuning (PEFT, LoRA), Huggingface or Unsloth
• Advanced training methods: Test-time training, (transductive) active learning, reinforcement learning
• High-throughput serving with vLLM
• Apply embedding models (e.g. SentenceTransformers), similarity/vector search or vector DB or ranking (e.g. LlamaIndex, Faiss, LangChain)
• Generate and filter synthetic data, clustering
• Detect hallucinations
• Evaluate chatbot models (Rouge, BLEU, F1-Score, Recall, Precision)
• Visualization of experiments (matplotlib)

AI-driven automation of purchasing on a printing company's website, including selecting delivery times, order options, ordering, payment, and uploading print data from the Lumiz Cloud.

• Consulting and interim management in the automotive industry, healthcare and other sectors focusing on program management, production launches, plant management, supplier development, cost reduction, process optimization, AI implementation and AI-driven process automation.

• Key projects:

• Concept for performance documentation of a pain therapy department with interdisciplinary input and reconciliation with OPS catalogue (Feb 2026).

• Complete billing system for a medical practice with multi-tenant support, state machine workflow, ZUGFeRD/EN16931-compliant invoice generation, OpenEMR integration; implemented with ~4 200 LOC Python, ~1 700 LOC SQL, 30+ endpoints, 13 DB migrations (Dec 2025 — Feb 2026).

• Clinical decision support system for inpatient pain therapy including automated processing of pain questionnaires, AI-generated therapy recommendations and longitudinal analysis; stack: FastAPI, Ollama, PyMuPDF, React + TypeScript, PostgreSQL, Docker (Sep 2025 — Mar 2026).

• End-to-end AI-powered transcription and documentation pipeline from physician-patient conversations with WhisperX transcription, five-stage LLM pipeline, review UI and self-hosted infrastructure; stack: FastAPI, WhisperX, Ollama, React + TypeScript, Docker (Aug 2025 — Feb 2026).

• AI-powered pipeline for document recognition and automated bank reconciliation processing >3 000 transactions and >4 000 documents with multi-model OCR, LLM-based document separation, web UI and loop-based prompt versioning; infrastructure: Docker Compose, PostgreSQL, Ollama LLM server (Aug 2025 — present).

• Supplier development for suspension parts including changes, new launches, relocations and bottleneck management for KTM (May 2023 — Nov 2024).

• Leading transfer of 150 serial production items after supplier plant closure for KTM (Jul 2022 — Jan 2024).

• Lean and process consultation for Swiss manufacturer of electro components at Von Roll Schweiz AG (Jun 2018 — Nov 2018).

• Ongoing back office support and process improvement for a private pain therapy & TCM practice including anesthesia protocols, website setup, digitalization and process optimization (May 2018 — May 2025).

• Acting plant manager for interior trim components in Hungary focusing on stabilization, tooling optimization, headcount reduction and backlog reduction at MAO Automotive GmbH & Co (Feb 2018 — Jun 2018).

• Start-up consulting and coaching for private pain therapy & TCM practice covering concept, business plan, financing, construction, setup and SOP implementation (May 2017 — Apr 2018).

• Program manager for critical suppliers DAG BR238 door trim leading supplier qualification, sampling, PPAP/EMPB tracking and process approvals at Megatech Industries Deutschland GmbH (Jul 2016 — Oct 2017).

• Interim COO and acting plant manager in aluminum profile processing automotive supplier with 350 employees and €80 M revenue, leading expansion in Slovakia at PWG Profilrollen-Werkzeugbau GmbH (Sep 2015 — May 2016).

• Leading labor and material efficiency program for headliner production focusing on process, material, labor and supplier cost reduction at Motus Headliner GmbH (Jan 2015 — Aug 2015).

• Leading development program for interior door trim DAG C292 including launch phase onsite in MS & AL at Toyota Boshoku America (Sep 2012 — Jan 2015).

• Leading cost down team for DAG C218 door panels at Toyota Boshoku Europe NV (Mar 2012 — Aug 2012).

• Sun visor development consulting for project team in Turkey including optimization, market analysis and design reviews at MARTUR FOMPAK (Jan 2012 — Dec 2014).

• Leading transfer of injection molding tools to new suppliers including sampling, assembly trials and PPAP/EMPB coordination at Toyota Boshoku Europe NV (Jan 2012 — Aug 2012).

• Program manager for BMW sun visors L7 Platform with production transfer and new development of F30 sun visor at Magna (Jul 2010 — Dec 2011).

• Evaluation and analysis of camera-based traffic light and sign recognition system on various LLM-based autonomous driving systems (LMDrive, BEVDriver)
• Implemented VLM based traffic notice instruction generation unit for closed-loop autonomous driving system which alerts driver in unforeseen driving incidents
• Developed independent LLM-based local chatbot with Llama, DeepSeek and Qwen including MLflow evaluation framework

• Further training in Snowflake and Google Looker
• Working with LLMs: local models (Llama, Mistral, Gemma, Phi, Qwen, DeepSeek, Bitnet, Flux, Whisper), OpenAI API, frontends (ollama, openwebui, loacalai)
• Inference methods: llama.cpp, vLLM, transformer
• Quantization, benchmarking, prompting
• LLM Agents (Tool/Function Calling, LangChain, LangGraph, MCP)
• Topics: attention, reasoning, chain of thoughts, RAG, GraphRAG, mlflow
• Cloud hosted: ChatGPT, Claude, Gemini

• Built a hybrid semantic and keyword search and LLM-based requirement extraction from conversational queries, boosting search accuracy by 85%, cutting zero-result searches by 70%, and reducing search time by 60%.
• Deployed a production-ready API with monitoring dashboards over 100K+ products, keeping response times under 2s and reducing customer search-to-purchase time by 40%.
• Technologies: Python, BGE-M3, Qwen2.5, FastAPI, Qdrant, Meilisearch, Docker, Prometheus, vLLM.