Large Language Model Experts in Stuttgart

• Task Engineering: Evaluating Coding Agents.
• Asset Creation: Building realistic task environments using datasets and files in German. Crucially, these assets must remain in the target language to genuinely measure multilingual handling.
• Prompting & Translation: finding failure points where AI does not work, in German.
• Implementation & Verification: Supporting the development of robust solutions (reference implementations) and write highly reliable, deterministic verifier scripts (using rubric-based judging only when strictly necessary).
• Calibration & Execution: Analyze execution logs and calibrate task difficulty (Easy to Very Hard) using standard Terminal-Bench run configurations against various model tiers (Haiku, Opus).
• Quality Assurance: Participation in a rigorous, 4-layer human quality control process (creation, human review, calibration review, and audit) alongside automated LLM-based checks to ensure fairness, grammatical accuracy, and benchmark integrity.
• Linguistic Review: Reviewing AI benchmark tasks across Hindi, Arabic, Japanese, Chinese, Czech and Turkish.

Methodology: Agile project management

Goal: Digitization and implementation of CCD2 requirements within the Consumer Finance systems.

Analysis, identification, and assessment of business requirements, as well as their transfer into technical implementation requirements. Coordination and status reporting of the various requirements. Documentation and process management using UML and BMPN. Cross-functional stakeholder management, UAT testing, data management, API architecture management.

Tools: Jira, Confluence, Visual Code, Sql Developer, Enterprise Architect, MsOffice, Bruno, ALM, Camunda, Figma

Designed and developed an AI-powered automation platform using n8n to analyze social media niches, identify target audiences, and automate marketing strategy generation. The solution combined AI agents, workflow orchestration, and data analysis to automate research processes and generate data-driven insights.

• Designed and implemented complex automation workflows using n8n
• Developed AI-powered analysis agents for market and audience research
• Integrated multiple APIs and AI services into automated workflows
• Built automated market, competitor, and target audience analysis pipelines
• Leveraged Large Language Models (LLMs) for information summarization, classification, and prioritization
• Containerized and deployed the platform using Docker

Technologies: n8n, AI Agents, OpenAI APIs, Prompt Engineering, LLMs, Docker, Linux, REST APIs, Webhooks

• Go-to-market and campaigns for SAP-related software solutions from positioning to execution.
• Buyer’s guides and sales tools,
• DACH and UK including tracking and attribution,
• Selection and management of the performance agency,
• Content concept for the DSAG Annual Congress trade fair booth,
• quality assurance of SAP-related content.
• Evaluation of two marketing platforms as a decision basis for management, implementation in HubSpot.

Two complementary tracks: (a) building my own AI products and open-source tools under the Neckarshore AI brand, with a focus on AI-supported multi-agent development processes and documentation automation; (b) freelance support for industrial, logistics, and financial companies in IT-related transformation projects.

Current product portfolio under the Neckarshore AI brand:

• Omnopsis Documenter — AI-powered documentation engine: generates compliance, technical, and release documentation from Git, Jira, and Confluence. 500+ automated tests, 100+ API endpoints, RBAC, monitoring stack.
• TrustScope — trust analysis for public GitHub repos: deterministic three-pillar report (Security & Supply Chain, Governance, Community) based on OpenSSF Scorecard, with constructive fix suggestions instead of a single misleading score.
• md-viewer — dependency-free Markdown viewer for macOS (Finder Quick Action) and web, split view of rendered document and raw text.
• Obsidian Vault Autopilot — AI-supported vault automation for Obsidian × Claude Code: sorts the inbox, renames notes, enriches frontmatter.
• Phonesis Voice Bank — voice archive for families, hospices, and cultural archives (guided recordings, inheritable archive, optional voice cloning); currently in the GDPR compliance phase (Art. 9, biometric data).
• Multi-Agent Development Process — 12+ specialized agents (architecture, implementation, security, refactoring, marketing) with structured handoff protocol, parallel execution, and automated quality assurance — powers the entire product development.
• Freelance consulting: Technical Product Ownership for production, logistics, and compliance/AML systems; steering international rollouts including test, cutover, and change management.

Stack: NestJS, TypeScript, PostgreSQL, Redis/BullMQ, Next.js, Claude Code (Opus), Docker, GitHub Actions, Vercel

• Situation: Increasing demand for privacy-compliant AI solutions for clients in the KRITIS and mid-market sector that need to analyze sensitive media content (audio, video, documents) without sending data to public cloud LLMs.

• Task: Design, deployment, and secure operation of a fully self-hosted AI infrastructure including a custom-built digital management platform for automated media analysis.

• Action: Architected and implemented a multi-tier platform on hardened Proxmox infrastructure with frontend (Nuxt 3, Vue 3, TypeScript, Tailwind 4), backend (Laravel 13, PHP 8.4, Sanctum), data storage (PostgreSQL 16, MongoDB 7), caching/queuing (Redis 7, Laravel Queue), AI workers (Python 3.11, Whisper, DeepFace, Librosa), scheduling (Laravel Scheduler/Cron), and local LLMs (Gemma, DeepSeek, Qwen, Mistral, LLaMA, Phi) via OpenWebUI with segmented network access, API hardening, and audit logging following BSI recommendations.

• Result: Fully GDPR-compliant, on-premises AI platform with zero data leakage to third parties.

• Task: Overall responsibility as an external Head of Cyber Security / CISO-as-a-Service for the design, implementation, and continuous improvement of ISMS according to ISO 27001, BSI IT-Grundschutz, and NIS2.

• Action: Built and managed Cyber Defense Centers (CDC) with SOC operations, integrated SIEM solutions (Splunk, Graylog), established risk-based vulnerability management (Qualys, Nessus, OpenVAS), and conducted regular infrastructure, application, and physical penetration tests.

• Result: Audit-ready ISMS for multiple clients and a 60% reduction in critical vulnerabilities within 90 days.

• Task: Design and execution of NIS2 assessments and operational roll-out plans for KRITIS operators.

• Action: Developed an online assessment tool for automated identification of individual weakness profiles, implemented ISMS optimizations, penetration testing, awareness programs, GRC suite deployment, and delivered C-level presentations.

• Result: Accelerated the consulting process by 50% and successfully prepared multiple clients for NIS2 compliance.

• Task: Incident commander for crisis response, forensics, and business recovery in ransomware attacks and APT campaigns.

• Action: Coordinated with state and federal police (LKA, BKA), performed forensic analysis (OSForensics, Wireshark, Kali Linux), executed disaster recovery and BCM strategies, and developed BTC extortion response strategies.

• Result: 100% recovery rate within defined RTO windows and sustainable post-incident security architectures.

• Action: Planned, built, and operated a hardened multi-VM infrastructure (Proxmox, 15+ VMs) with web and mail servers, Graylog, OPNsense firewalls, CRM/ERP and LLM instances, network segmentation, DDoS mitigation, automated patch management, and backup strategies.

• Result: >99.5% uptime over 20+ years and zero compromises.

• Action: Designed coordinated phishing campaigns with five levels of difficulty, developed e-trainings and webinars in a PDCA cycle, and led red and blue teams.

• Result: Phishing click rate reduced from 35% to under 5% within three campaign cycles.

• Development of NLP and LLM POCs for use in manufacturing companies
• Application of advanced machine learning algorithms to analyze production data and develop customized data pipelines for quality assurance
• Design and management of the IAO foundational seminar on AI in industry, including practical training modules

• Developed and deployed a deep learning model for automated keypoint detection in headlamp light distributions.
• Prepared and processed datasets, and selected VGG16 after benchmarking CNN architectures for the best accuracy efficiency trade-off.
• Delivered a Flask REST API, containerized with Docker, and integrated the solution into an existing internal system, enabling automated and efficient evaluation of headlamp designs.

• Lead Product Owner for 4 development teams
• Leading and coordinating a greenfield project with parallel implementation of core components by independent teams; managing dependencies and resources
• Establishing a data lakehouse approach, including analysis of data volumes and future requirements as part of a cloud migration (best-of-breed approach)
• Responsible for requirements analysis, selection, and piloting of a LIMS/ELN system, supported by advising decision-makers and managing external vendors
• Introducing and managing an OpenWeb UI and Azure OpenAI-based RAG system to support knowledge extraction and data-driven analyses
• Setting up, configuring, and managing Jira projects, as well as developing project-specific workflows and automations
• Implementing classic Scrum processes with all ceremonies and taking on the Scrum Master role for all involved teams
• Assisting in hiring through interviews and assessments from a product owner's perspective
• Making key architectural decisions, including selecting the platform for the data lakehouse (Databricks) and the strategic integration of LIMS and analytics platforms

• Founded Latence to commercialise runtime safety patterns from HALO as a deployable product.
• Built end-to-end as single technical founder with open-source stack on NVIDIA ecosystem.
• Developed TRACE: real-time safety layer for knowledge agents and RAG pipelines with groundedness scoring, prompt-attack detection, GDPR redaction, context compression, audit-ready traces.
• Developed vLLM Factory: production inference framework on vLLM with custom Triton kernels and 12 parity-validated plugin models, achieving up to 11.7× throughput vs vanilla PyTorch.
• Developed ColSearch: single-node multi-vector late-interaction retrieval engine with Rust SIMD and fused CUDA, achieving 3.12× FastPlaid geomean QPS on BEIR-8 and a 1.58-bit quantized lane 6.4× smaller than FP16.
• Developed llm-opt: LLM compression research framework with hierarchical importance, structured pruning, tabu search, knowledge distillation.

• Product owner and Executive Sponsor for AI Assistant trained on a knowledge base of past solution designs, RFP documents, and current product documentation
• Used by close to 100 global architects and engineers in pre-sales and post-sales
• Generates full solution documents with requirement driven architecture decisions, solution overview, diagrams, bill of material, roles and responsibility matrix
• Technologies: AI, LLM, Chat Bot, RAG, Agentic AI, Vector Databases, Public Cloud

• Teaching German literature and linguistics

• Consulting on lean, logistics, AI applications and process optimization
• Customized concepts for digitalization, change & transformation
• Adaptive training and workshops for professionals and managers

• 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