Open to onsite, hybrid & remote roles
Senior Data EngineerBerlin, Germany
I build lakehouse platforms, streaming pipelines, LLM-powered data products and data teams can trust.
5+ years turning raw, complex data into clean, usable data products on Databricks, Spark, Kafka and the cloud. I started out in data science, so I build pipelines with the people who use the data in mind: accurate, well modelled and easy to access.
01Skills
Production-tested across Databricks lakehouses, Kafka streams, AWS and Azure: the everyday tools behind reliable, well-governed data.
Used in productionCertification-based
Where most of my days go: taking Databricks from proof of concept to production, from incremental ingestion to governed, analytics-ready tables.
The everyday toolkit behind every pipeline.
Real-time analytics on a Kafka backbone of roughly 5–10 TB of raw events a day.
Scheduling, ingestion and alerting, with incremental loads by default.
Versioned, tested and deployed as code.
AWS and Azure in production; Google Cloud and some AWS services through certification.
AWS
Azure
Google Cloud
LLMs where they pay off, people where judgement matters, and agentic tooling for the engineering itself.
Clear contracts in, trustworthy tables out.
From source systems to the dashboards people open every morning.
Listening first, then building what the data's users actually need.
02Experience
Four roles, one thread: making data dependable for the people who rely on it.
Featured project · MARKT-PILOT
Industrial spare-parts data arrives messy: manufacturer names written many different ways, duplicate parts, gaps. I designed and built the end-to-end pipeline that cleans, deduplicates and verifies it with LLMs and web research, and routes what it can’t match to a review app whose decisions flow back into the pipeline.
of spare parts now matched automatically, without manual review
of manufacturer names folded into their canonical form
of parts deduplicated
01 · Ingest
Bronze
MongoDB → Airbyte → S3, read by Databricks as an external location
02 · Refine
Silver
Lakeflow Spark Declarative Pipelines: CDC-based incremental loads plus materialized views
03 · Match
LLM matching
Manufacturer cleaning, deduplication and verification with ai_query and SerpAPI web research
Loop · Review
Human in the loop
Unmatched parts go to a review app; every decision feeds back into matching
04 · Model
Gold
Star schema with data contracts and automated quality checks
05 · Serve
Analytics
Reusable data products for internal analytics and executive dashboards
MARKT-PILOT
I own the silver layer of a Databricks medallion lakehouse for industrial spare-parts data, and the LLM pipeline that matches it.
Bentego · Consultant for Vodafone
Real-time streaming and batch pipelines for Vodafone, on a Kafka backbone ingesting roughly 5–10 TB of raw events a day.
Borusan Otomotiv
Enterprise reporting on Microsoft Azure, from pipelines to warehouse to dashboards.
Borusan Otomotiv
Where it started: on the consuming side of data pipelines, which is why I build them with their users in mind.
03Credentials
Two professional data engineering certifications, a computer engineering degree and three languages.
Associate level
Amazon Web Services
Verify on CredlyProfessional level
Google Cloud
Verify on CredlyIzmir University of Economics · Izmir, Turkey
04Writing
I write about what I learn on the job: certifications, cloud security, data governance, orchestration and the basics of machine learning.
Drag or scroll sideways
All articles on Medium05Beyond work
Breathe in, breathe out, repeat.
The best way to get to know a city.
New places, new perspectives.
Turning what I learn into articles.
06Contact
I’m open to Data Engineer and Senior Data Engineer roles, whether onsite, hybrid or remote. I’m based in Berlin, and email is the quickest way to reach me.