Freelancer profile translated to English.
Back to original languageAbout Dennis
Are you planning a GenAI or ML project and looking for someone who can not only outline your AI system but also build it production-ready? That's exactly my focus.
As an AI Architect & ML/GenAI Engineer with over 6 years of hands-on experience, I help companies transition AI systems from proof-of-concept to robust, operational solutions. My core expertise lies in the two areas where most AI projects fail today: production-ready LLM engineering and cleanly set up MLOps pipelines.
What makes me unique: I'm not just an engineer; I was also most recently a co-founder of a funded AI startup in the AIOps space. I personally worked daily on a system that is in production with pilot customers. I bring this practical experience, including all the real-world compromises (latency, cost, scaling, compliance, maintainability), directly to your projects.
Typical services I offer:
LLM and GenAI Engineering: Conception and implementation of RAG architectures, fine-tuning European open-source models via PEFT/LoRA, multi-provider LLM gateways, MCP tool integration, LLM observability, and guardrail strategies.
MLOps and ML Lifecycle: Building experiment tracking, model versioning, and workflow orchestration with MLflow, Kubeflow, and Airflow. Reproducibility as a central principle.
Anomaly Detection and Time Series ML: From classic methods to LSTM autoencoders and Transformer architectures, applicable to infrastructure, production, or business data.
General Data Science topics for predictions or classification.
Consulting and Sparring: Architecture reviews, make-or-buy decisions for AI tooling, technical sparring partnership for CTOs and Tech Leads.
German
Native or bilingual
Spanish
Fluent
English
Conversational
Can work on-site
Frankfurt am Main (up to 20km)
Experience
- aprevisCo-Founder & Head of AI/MLSOFTWARE PUBLISHINGFebruary 2025 - Today (1 year and 6 months)Darmstadt, GermanyCo-founder with primary technical responsibility. Conception and development of the ML/DL pipeline for anomaly detection on Oracle system metrics. Development, training, and validation of models (Transformer-based, LSTM autoencoders, Isolation Forest, One-Class SVM). Architecture of the root-cause analysis pipeline with LLM-supported hypothesis generation. Fine-tuning European open-source models via PEFT/LoRA. Multi-provider LLM gateway, RAG integration, MCP tool connection, backend in Python.Qualifications:Python, FastAPI, TensorFlow, PyTorch, Hugging Face, MLflow, Kubeflow, Airflow, LangChain, LiteLLM, Langfuse, LangSmith, Pinecone, Qdrant, RAG, Fine-Tuning, PEFT, LoRA, MCP, Mistral, DeepSeek, OpenAI API, LLM-Observability, Anomaly Detection, LSTM Autoencoders, Isolation Forest, Transformer, Time Series, Oracle, Vue.js, Linux, Data Science
- Premium-AutomobilherstellerLead Python Developer & Data ScientistAUTOMOBILEMarch 2024 - February 2025 (11 months)München, GermanyTechnical lead of a four-person development team (Lead Python Developer) for the development of a platform for automated component assessment and quotation for a German premium automobile manufacturer. Backend architecture in Python/FastAPI, connection of internal component data via GraphQL, component classification with Azure OpenAI and embedding-based matching, modeling of the Bill of Materials in a graph database for variant management and impact analysis. Consulting and upskilling of business departments on Data Science topics such as price prediction and model deployment.Qualifications:Python, FastAPI, Azure OpenAI, GPT-4, PostgreSQL, Neo4j, Cypher, GraphQL, React, Jupyter, Embedding, Vector Search, Bill of Materials, Variant Management, Supply Chain, Predictive Modeling, Machine Learning, Tech Lead, Architecture, GenAI, LLM Integration, Training, Python Training, Stakeholder Management, Data Science
- CapgeminiMLOps ResearcherDIGITAL AND ITJanuary 2024 - February 2024 (1 month)Frankfurt am Main, GermanyInternal strategic evaluation of MLOps frameworks for standardization. Comparative assessment of MLflow, Kubeflow, and Flyte based on weighted criteria. Development of proof-of-concepts to validate workflow capabilities.Qualifications:MLflow, Kubeflow, Flyte, MLOps, Framework Evaluation, Proof-of-Concept, Workflow Orchestration, Experiment Tracking, Model Versioning, Model Registry, Reproducibility, Python, ML Lifecycle, Tech Strategy, Consulting
Recommendations
Be the first to recommend Dennis
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Master of Science - Data ScienceHochschule Darmstadt (dual with ORDIX AG)2021Dualer Master mit Praxisphasen bei einem Datenbank- und IT-Beratungsunternehmen. Forschungsschwerpunkte: Deep Learning, Generative AI, Reinforcement Learning, Time-Series-Analyse, Data Science, Data Analytics, Data Visualization. Forschungsarbeit zu Reward-Hacking-Effekten im Reinforcement Learning. Masterarbeit zur Generierung interaktiver elektronischer Musik mit selbst implementierten Autoencoder- und GAN-Architekturen in TensorFlow, Training auf Google-Cloud-GPUs. Note 1,3 für die Masterarbeit, Gesamtnote 1,2. Abschluss Dezember 2021.
- Bachelor of EngineeringDuale Hochschule Baden-Württemberg (DHBW) Stuttgart2018Duales Studium mit Schwerpunkt Produktion und Logistik bei einem deutschen Automobilzulieferer. Praxisphasen in der Produktionsplanung, in Make-or-Buy-Analysen sowie bei Auslandseinsätzen in Werken in Ungarn und Polen. Fünftes Semester als Auslandssemester an der Edinburgh Napier University in Schottland. Bachelorarbeit zur Optimierung der Produktionsplanung mit der SAP-Long-Term-Planning-Funktionalität. Gesamtnote 1,9. Abschluss September 2018, gefolgt von 9 Monaten Festanstellung als Produktionsplaner.
Certifications
- TensorFlow Developer CertificateGoogle
- IBM Data Science Professional CertificateCoursera2019