Plotly Dash Experts in Germany

Architected and maintained data products for renewable energy operations, covering wind turbine, grid-meter, and weather data. Built scalable ETL/ELT pipelines in Azure Databricks using Delta Lake (bronze/silver/gold layers) and processed data in various formats, including structured and semi-structured data. Contributed to a data quality framework supporting table and column documentation, outlier detection, and completeness metrics across all datasets within a data product. In addition, implemented a DORA KPI Databricks dashboard used across all data products. Optimized CI/CD processes in Azure DevOps to streamline deployment across development, test, and production environments.

Technology stack: Azure Databricks, PySpark, SQL, Delta Lake, Unity Catalog, Azure Data Lake, APIs, Dremio, Azure DevOps, YAML, Git, Databricks Workflows, Application Insights, Terraform, OpenAI API, Codex, LLM-assisted workflows

• Building an AI & Data Consultancy Practice with the goal of helping European companies adopt Artificial Intelligence and modern data platforms
• End-to-end further development of a production system using modified coding agents (OpenCode). Tech stack: Kubernetes, Argo, Keycloak, Typescript, Grafana, GitOps, DevOps, Playwright
• Internal research project on the use of coding agents in the field of mathematical logic for creating formal models. Use of Cursor IDE and Codex, Codex CLI. Architecture design, quality control and refactoring, as well as writing code and tests. Repository (open source) available pre-launch
• Research on the role of mathematical logic as a formal language that connects IT and AI with business processes
• Project lead for collecting and deploying parking recommendations for rail vehicles with significant savings potential based on real-time data in a mobility and transport company
• Project lead for collecting and distributing process measurement points for real-time control in a mobility and transport company
• Deputy application owner for an app used for communication in the dispatching and provision of rail vehicles

• Taught German language and mathematics to children aged 4-16, providing homework assistance and tutoring
• Supported startups in UX, MVP development, and Lean Startup methodology
• Assisted with branding, communication, and design to strengthen market presence
• Maintained website and Venture Lab app; coordinated and ran events
• Developed presentations to support internal and external communications

End-to-end responsibility for all products - concept, architecture, development, and production operation as the sole developer; in addition, customer meetings, proposals, and marketing.

Underwriting Copilot - AI assistant for industrial insurance (in production at customer sites)

• Supports underwriters in analyzing industrial insurance submissions - in production use at an industrial insurer.
• Framework-independent RAG architecture with Hybrid Search (BM25 + pgvector) across large, mixed document sets.
• 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 at customer sites)

• Automatically captures and analyzes menu 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.
• Full CI/CD pipelines (GitHub Actions), production cloud deployment, interactive dashboards (Dash).

Kaiki GEO Atlas - GEO platform (in production at customer sites)

• 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 extra LLM cost).

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 progress)

BankingGPT - AI assistant for complaint management in cooperative banking

• Security architecture at the core: no AI draft reaches the customer without human approval - the approval decision is in auditable code, not in the language model (monotonic: the model may escalate, never downgrade).
• Real agentic building blocks, each with its own boundary: the model chooses tools itself through an MCP server (read-only, allowlist, capped, fail-safe); sensitive cases are handed off via an open A2A protocol (JSON-RPC, Agent Card, message/send/tasks/get; client implemented by me) to a separate specialist agent (securities/law), which never lowers the review requirement (pinned by test).
• Evaluation-driven over ten analysis rounds; uncovered a security flaw through independent review and blind tests that nine automated runs had missed.
• Voice AI frontend, responding live: covered cases are answered in the conversation, sensitive ones escalate before generation; response latency < 7 s measured (local GPU STT/TTS).

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

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

• Genuinely agentic on LangGraph: ReAct agent with four tools and conversation memory - the model decides on its own whether to use the manual (RAG, Chroma), a knowledge graph (GraphRAG, Neo4j/Cypher - decodes warning lights), or a workshop/booking service.
• Human-in-the-Loop before the irreversible action: before every appointment booking, the graph pauses (interrupt) and gets the driver's explicit confirmation - the same approval-before-action discipline as in BankingGPT, in a different framework.
• Eval as CI gate: a three-part scorecard (RAGAS grounding + deterministic tool-routing accuracy + DeepEval safety: does the answer mention the warning first when there is 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.

• Bridged business and engineering for enterprise finance clients, designing data products and cloud pipelines in Python, SQL Server, SAP Datasphere, and Tagetik
• Conceived, built, and containerised a Python/FastAPI universal connector that syncs SAP S/4HANA and other SQL/NoSQL sources to Tagetik, deployed on Google Cloud Run and Microsoft Azure, cutting a critical 90-minute data load to approximately 80 seconds (65× faster)
• Architected a medallion-layer SQL Server warehouse ingesting approximately 500 GB/day from 11 ERP instances, automating daily refreshes (full load under 6 minutes) and freeing 20–30 finance staff from days of manual data consolidation
• Led cross-functional workshops to design enterprise EPM target architecture for a leading Southeast-Asian telecom (CAPEX, OPEX, revenue), translating requirements into data-model specifications and integration blueprints now being built by the client’s implementation team
• Delivered selected projects including a consolidated data & reporting warehouse for a global manufacturer (10 k+ employees), NFI reporting for an international management & technology consultancy, and CAPEX/OPEX planning for a Southeast-Asian telecom (20 k+ employees)

• Architect AI systems across build-vs-buy layers; guide clients on technology selection, evaluation, integration patterns, and governance to reduce risk and time-to-value.
• Implement Azure/Snowflake solutions (RAG pipelines, chatbots, data agents) including ingestion, retrieval, orchestration, and monitoring.
• Partner with stakeholders to translate business needs into deployable AI roadmaps and reference architectures; align with existing data platforms and security controls.

• Building the company with a focus on connectivity for IIoT and Industry 4.0, iRPA/process automation, advanced robotics and smart systems, sensors and services
• Project management and software architecture for IoT gateway development (since 2020) with protocol translation, IT/OT convergence and GRC
• Developing RPA bots for automating and monitoring industrial processes with an agent-based AI approach (since 2020)
• Implementing unsupervised clustering and anomaly detection for time series data in big data streaming pipelines (since 2021)
• Introducing a Docker-based release train for OTA updates with DevSecOps and CI/CD (since 2018)

• Built an AI Research Assistant with RAG, LangChain, LangGraph, and OpenAI LLMs integrated with vector search; cut research time by 30%.
• Designed custom retrieval workflows with LlamaIndex, building a ReAct-style agent for dynamic chunking; improved query accuracy by 18%.
• Researched and optimized embedding strategies, reducing retrieval cost/query by 15%.
• Developed RAG evaluation frameworks using RAGAS and Langsmith with custom datasets; improved coverage by 40%.
• Fine-tuned LLMs (LLaMA 2 on Vertex AI with custom inference containers, dynamic batching, and quantization); reduced inference latency by 25%.
• Integrated AI agents in LangGraph with short-term & long-term memory (Mem0); increased task completion rate by 20%.
• Created schema-aware synthetic data generators; fine-tuned downstream models achieving +12% F1 score.

• Built advanced expertise in designing and deploying Model Context Protocol (MCP) infrastructures to enhance AI agent capabilities for an NGO.
• Developed and maintained MCP servers (e.g., using FastMCP) exposing external tools and data sources to LLMs via a standardized API (JSOB).
• Enabled seamless tool integration by adhering an MCP client-server architecture, allowing AI hosts or agents to dynamically discover and invoke tools, access resources, and leverage custom prompts (File Storage, SQL DBs, weaviate, chroma).
• Automated deployment (CI/CD) in AWS with IaC (CDK) and Jenkins (scheduler).
• Conducted comparative model evaluation due to German language needs (Hugging Face over Anthropic, Coehere, others) and AWS Bedrock.
• Evaluated legal and regulatory aspects, ensuring the MCP-based solutions complied with emerging EU AI Act requirements, such as risk classification, transparency, documentation, and responsible tool provisioning within the European context.

• Acted as the primary bridge between Legal, Engineering, and Business units to ensure zero compliance violations while maintaining product velocity.

• Led the development of a GDPR-compliant AI aggregator platform, managing a roadmap that balances legal constraints with aggressive feature delivery.

• Scaled the engineering team from 4 to 9 developers, establishing hiring protocols and technical onboarding processes to support rapid product iteration.

• Boosted the development process by introducing structured sprint cycles and backlog refinement, resulting in a 20% reduction in feature delivery time.

• Architected and prototyped agentic AI workflows with n8n and RAG pipelines on Langchain.

• Supported the development of AI-powered voice assistants for recruitment automation.
• Helped optimize speech recognition models and conversational AI workflows.

• Deployed and optimized Vision-Language-Action (VLA) and diffusion policy models on NVIDIA Jetson Orin and Jetson Thor, meeting real-time inference latency targets for humanoid robot control loops.
• Built TensorRT engine pipelines (PyTorch → ONNX → TensorRT) with INT8/FP8 post-training quantization, calibration dataset design, and quantization-aware validation, reducing inference memory footprint by over 3× on Jetson without accuracy regression.
• Developed custom CUDA C++ plugins and CUDA Graphs for latency-deterministic, real-time policy execution – meeting hard runtime and memory constraints on embedded GPU targets.
• Developed an inference engine for VLA models on top of llama.cpp bringing different VLA policies under single runtime, packaging each as a single self-contained GGUF that needs no Python or PyTorch.
• Profiled and tuned GPU execution using NVIDIA Nsight Systems and Nsight Compute, identifying CUDA kernel bottlenecks, memory bandwidth saturation, and SM occupancy issues across Jetson Orin and Thor compute profiles for cross-layer performance optimization.