Building reliable data pipelines and platforms for analytics and machine learning.

I am a Data Engineer and ML Engineer with a research background in machine learning and reinforcement learning.

During more than three years of postdoctoral research at TH Ingolstadt / AImotion Bavaria, I worked on reinforcement learning agents, simulation environments, forecasting models, and the supporting infrastructure required to run reproducible ML experiments.

I later completed an intensive 720-hour Data Engineering Bootcamp, where I focused on building production-style batch and streaming data pipelines, cloud data platforms, orchestration workflows, data-quality layers, and analytics-ready models.

Today, I combine both sides of that background: data engineering for reliable data foundations and ML/AI experience for systems built around real analytical and modeling needs.

• 🔭 Focus: Data Engineering · ML Engineering · Applied AI
• ⚙️ Engineering: Batch & streaming pipelines · Data modeling · Orchestration · Data quality · Cloud
• 🧪 ML/AI: Reinforcement learning · Forecasting · PyTorch · Reproducible ML workflows
• 📍 Based in: Ingolstadt, Germany

End-to-end batch analytics platform processing 101.7M Deutsche Bahn operational observations.

Hugging Face → S3 Bronze → Great Expectations → Redshift Serverless → dbt → Metabase

• Medallion architecture with Bronze, Silver, and Gold layers
• Monthly pipeline orchestrated end to end with Apache Airflow
• Data-quality gates with Great Expectations, pytest, and dbt tests
• Deduplication and modeling of poll-and-snapshot railway data
• Analytics-ready marts for station, service, delay, on-time, and cancellation performance
• Interactive Metabase dashboard over Gold models

Stack: Python · SQL · Airflow · dbt · AWS S3 · Redshift Serverless · Great Expectations · Metabase · Docker

Real-time pipeline streaming live Wikimedia edits through Kafka into MongoDB and a live dashboard.

Wikimedia SSE → Python Producer → Kafka → ksqlDB → MongoDB → Streamlit

• Streams live Wikimedia RecentChanges events
• Computes one-minute aggregations with ksqlDB
• Persists raw and aggregated events in MongoDB
• Uses idempotent consumers to avoid duplicates during restart or reprocessing
• Visualizes live activity in a Streamlit dashboard
• Runs locally as a reproducible Docker Compose stack

Stack: Python · Kafka · ksqlDB · MongoDB · Streamlit · Docker Compose · pytest

Production-style batch pipeline from Brazilian e-commerce CSV data to analytics-ready Snowflake models.

CSV → Airflow → Snowflake → dbt → Tests

• Loads source datasets into Snowflake
• Models data through staging → intermediate → marts
• Orchestrates the complete workflow with a daily Airflow DAG
• Runs dbt dependency, compile, transformation, and test steps automatically
• Fully containerized with Docker

Stack: Python · SQL · Airflow · dbt · Snowflake · Docker · uv

Data Engineering & Analytics

Cloud, Infrastructure & DevOps

Machine Learning

• Data Engineering Bootcamp — 720 hours
  Hands-on training in Python, SQL, AWS, Airflow, dbt, Spark, Kafka, Docker, Terraform, data modeling, orchestration, and cloud data architectures.

• Postdoctoral Research — ML/AI, TH Ingolstadt / AImotion Bavaria
  Reinforcement learning, simulation environments, forecasting, and ML infrastructure for gas-network safety.

• PhD — Modeling & Control of Multimodal Transport Systems
  Cadi Ayyad University, Morocco.

• Research published at ECAI, ECML-PKDD, and IEEE venues.

Open to Data Engineer and ML Engineer opportunities.
Feel free to connect with me on LinkedIn or reach out by email.