Data Engineer
I build data platforms the way a strong player builds a position: a sound opening, every piece developed with purpose, and no loose pawns in production. Batch, streaming and lakehouse pipelines that stay calm under time pressure.
—Starting position
Who is at the board — ! best by test
I am a Data Engineer based in Tychy, Poland, with eight years of experience turning scattered operational data into reliable, well-modelled platforms. My work lives between the source systems and the people who need answers: ingestion, transformation, orchestration, storage and the quality checks that make all of it trustworthy.
I have built real-time event pipelines for payments, migrated legacy Hadoop warehouses into modern lakehouses, and designed dbt projects that analysts actually enjoy working in. I care about idempotent jobs, clear data contracts, sensible costs and documentation that survives the person who wrote it.
Like any good opening, the first few decisions in a data platform shape everything that follows. I like getting those decisions right, and then playing the long game patiently.
Opening principles
One governed source of truth with explicit data contracts, so every downstream team plays from the same position.
Automate early: CI/CD for pipelines, infrastructure as code, tests that run before anything reaches production.
Security, privacy and lineage are built in from move one, not patched in after the attack.
Every skill has a role on the board — select a piece
Moves one square at a time — and everything else exists to protect it.
The decisions that the whole platform depends on: dimensional and Data Vault modelling, medallion layering, data contracts, partitioning strategy and governance. Slow, deliberate and never left exposed.
Moves in any direction, any distance — the most versatile piece I own.
Distributed batch processing, custom connectors, API ingestion, tooling and libraries. PySpark jobs tuned for skew, shuffle and spill; clean, typed and tested Python everywhere else.
Straight lines, open files, decisive power in the endgame.
Analytical SQL, window functions, query plans and cost tuning across cloud warehouses. Transformations versioned and tested with dbt, served to BI tools through a consistent semantic layer.
Cuts diagonally across the board — events arrive from every angle.
Event-driven architectures with exactly-once semantics, change data capture from operational databases, windowed aggregations and low-latency feature serving.
The only piece that jumps — over dependencies, retries and backfills.
DAG design that is idempotent, observable and easy to backfill. Asset-based scheduling, SLAs, alerting and lineage so that when something breaks at 3 a.m., the fix is obvious.
Doubled rooks on an open file — infrastructure that backs up every attack.
Designing, provisioning and cost-controlling data infrastructure in the cloud: object storage, serverless compute, managed Spark and Kubernetes, all defined as code.
Selected studies — each line annotated, each result recorded
A real-time fraud-scoring pipeline for a card payments processor. Transactions stream through Kafka, Flink computes rolling behavioural features per card and merchant, and a Redis feature store serves them to the scoring model within milliseconds.
1. Debezium CDC → Kafka 2. Flink windowed features! 3. Redis serving 4. Delta Lake audit trail!!
Moved a logistics group's ageing on-premise Hadoop and Oracle warehouse to a Databricks lakehouse on Delta Lake, with zero-downtime dual-running, automated reconciliation and a table-by-table cutover plan.
1. Inventory & lineage map 2. Bronze/Silver/Gold layers 3. Row-level reconciliation! 4. Decommission Hadoop
Rebuilt a tangle of stored procedures into a modular dbt project on Snowflake with a semantic layer, so finance, product and operations finally agree on what "active customer" means.
1. Staging models 2. Tested marts! 3. Semantic metrics 4. Slim CI on every PR
A data observability layer that forces bad data to reveal itself: freshness, volume and schema checks wired into Airflow, lineage captured with OpenLineage, and incidents routed to the owning team before a dashboard ever shows a wrong number.
1. Great Expectations suites 2. OpenLineage graph 3. Anomaly detection! 4. Owner-routed alerts
The game record — career moves, with evaluation
| Move | Period | Position held | Eval |
|---|---|---|---|
| 4. | 2022 — Present |
Senior Data EngineerNordlicht Payments · Remote
|
+2.4 |
| 3. | 2020 — 2022 |
Data EngineerVistula Logistics Group · Katowice
|
+1.6 |
| 2. | 2018 — 2020 |
BI & ETL DeveloperSilesia Analytics Lab · Gliwice
|
+0.8 |
| 1. | 2014 — 2018 |
B.Sc. Computer ScienceSpecialisation in databases & distributed systems
|
= |
The king is safe — the board is open for your move
Your move.
Planning a new data platform, rescuing a fragile pipeline, or looking for a Data Engineer to join your team? Send a challenge and I'll reply within one working day.