AI / ML
Neural nets, deep learning, production ML
CEO & Managing Partner @ control-f GmbH. Machine learning engineer with 10+ years of data science and AI shipped at Porsche, Daimler, and Mercedes-Benz.

Exploring public data out of curiosity. Sources, observation dates and limitations are documented in each dashboard.
Small experiments, open questions, and things worth building just to find out. Notes from my projects in AI, engineering, and the world around me.
Research figureAn AI agent compared a DIE ZEIT reportage from the Strait of Hormuz with radar satellites, optical satellite images and ship data. Of 13 statements, 8 are correct, 5 are partly correct and none is false.
ScreenshotAn open-source game-agent benchmark with actual gameplay, measured reaction times and evidence you can inspect. Explore the first results and a visual catalog of game environments.
ScreenshotAn interactive view of global economic and market conditions, with country comparisons, historical context and visible source dates.
Four vectors. Numeric confidence from self-assessment against shipped production systems — not a LinkedIn endorsement list.
Neural nets, deep learning, production ML
Pipelines, warehouses, distributed compute
Full-stack systems, APIs, TypeScript first
Product, strategy, team building
Building an AI product studio: cutting-edge AI, data science, and proprietary systems for enterprise clients worldwide.
Tailor-made AI solutions for organizations. Seamless integration of ML into existing systems to optimize productivity and decision-making.
Vehicle data pipelines on the Porsche Data Platform (Cloudera/AWS + SAP DWC). Operationalized data views, quality-driven analytics, stakeholder alignment.
Led R&D on the next-generation Mercedes-Benz Infotainment System. Dynamic content platform integration and containerization for performance and scale.
Microservice + ML architectures across cloud and on-prem. Project Owner for a 1TB+ procurement savings prediction pipeline for Daimler's procurement org.
Coordinated AI-enabled software projects. NLP models for text analysis and classification; managed Azure infra for scalable hosting.
Freelance ML software development. Preprocessing and cleansing of large datasets, stakeholder-facing visualizations.
Facial-recognition room guidance system over multiple camera streams. Pathing algorithm and OAuth2-based LinkedIn crawler for embeddings.
Thesis: "A Deep-Learning Approach on Automotive Maneuver Detection." CAN-Bus preprocessing pipeline, classification across architectures, and VAE-based latent visualization. Earlier: live Jenkins analytics dashboard (d3.js + Flask).
Anomaly detection monitoring tool (SVM pre-classifier + Isolation Forest). Python-based IT process automation. Cloudera Hadoop test-build on-prem.
Dual-study program. Automated server installation, custom PostgreSQL databases, BI reporting, and change-management automation.
Daniel has a deep understanding of software architectures and knows how to build them from scratch. Great knowledge of cutting-edge technologies — and knows how and when to use them.
Constantly proves his great expertise in machine learning and cloud architecture. A proactive, solution-oriented colleague driven by creativity and curiosity for current trends.
A very good software developer with a broad knowledge of big data architectures. Always up to date — and great at finding creative solutions to the various obstacles along the way.
A very broad technical background combined with remarkable passion for data science. He pushed the products forward with an eye for detail and great curiosity for the latest trends.
A very inspiring personality, full of ideas and with a broad technical background. Daniel is able to deliver and prototype software products extremely quickly — he has internalized the agile mindset.
A creative mindset and a very special passion for the future of artificial intelligence. He is not only talking — he is doing things.
Extremely reliable and open-minded. Deep technical skills and practical experience in managing complex projects. Would I work with Daniel again? Yes, of course.
Third-party signal — talks, guest articles, panels. Every entry links a real, currently-reachable source, quoted straight and never paraphrased.