{ "version": "1", "data": { "cases": [ { "slug": "enyquant-2026", "client": "Enyquant", "role": "Platform Architect & end-to-end Data Engineer — Contract", "period": "2026 — present", "summary": "Built the company's cloud and data foundation from zero as its sole hands-on technology delivery owner, from architecture and code to deployment and daily operations.", "stack": [ "Azure, Alibaba Cloud", "Airflow, DuckDB, Polars, DuckLake", "Databricks, ADF", "Terraform + CDKTF (TypeScript)", "GitHub Actions + OIDC", "Python, SQL" ], "outcomes": [ "Delivered a five-year, country-wide electricity operations and market dataset to downstream SMEs within 60 days of starting the platform", "Defined and implemented point-in-time visibility semantics across nearly all datasets and model-training flows", "Re-platformed day-to-day development and operation from Databricks + ADF to a single-VM DuckDB architecture, reducing comparable monthly cost by about 95%", "Established a clear source of truth and non-conflicting team and personal rules for parallel AI-agent work" ] }, { "slug": "vodafoneziggo-snowflake", "client": "VodafoneZiggo", "role": "Data Engineer & Cloud Administrator — Freelance", "period": "2022 — 2024", "summary": "Worked inside a 100+ person Data Tribe on a shared 1PB+ core platform: evolving its existing ETL framework, supporting a hundreds-TB Oracle-to-Snowflake migration, and building a shared MLOps environment.", "stack": [ "Snowflake, Oracle", "AWS DMS (Full Load + CDC)", "Terraform, AWS CDK (TypeScript)", "Informatica PowerCenter, GitLab CI/CD", "Scala, Spark, Apache Oozie", "JupyterHub, MLflow" ], "outcomes": [ "Operated AWS DMS Full Load + CDC tasks for a hundreds-TB Oracle-to-Snowflake migration", "Embedded PowerCenter workflow conversion and early validation into CI/CD for at least 40 developers; feedback moved from hours or days to minutes", "Conservatively removed at least 40 hours of manual work per week across developers and the migration gatekeeper", "Built a shared JupyterHub + MLflow MLOps environment covering experimentation, registry, scheduling, and deployment for 4–5 data scientists" ] }, { "slug": "pvh-aws-datalake", "client": "PVH Corp (Tommy Hilfiger, Calvin Klein)", "role": "Lead / Senior Data Engineer — Freelance", "period": "2020 — 2022 & 2023 — 2025 (returned engagement)", "summary": "In a roughly ten-person core team without a dedicated Data Architect, served as the de facto architecture lead for the continued evolution of a 500+TB AWS data platform.", "stack": [ "AWS (S3, Glue, EMR, Lambda, Athena, Step Functions, EMR Serverless, DynamoDB)", "Spark, Airflow, dbt", "Terraform, AWS CDK", "GitLab CI/CD", "Azure Databricks, GCP BigQuery, Google Analytics" ], "outcomes": [ "Designed external integrations including Adobe, Salesforce, and SAP during the first engagement", "Made ETL idempotent and configuration-driven; improved observability, data quality, and cross-time-zone scheduling", "Enabled 60+ governed dashboards; a representative end-to-end delivery path fell from 4–6 weeks to about 1–2 hours", "Rebuilt an approved PII re-identification workflow from hourly batch and always-on infrastructure to event-driven serverless; typical latency fell from about 2 hours to 5–10 minutes and related AWS cost by about 90%", "Participated in Azure Databricks solution reviews and supported migration technical issues; did not own the migration" ] }, { "slug": "fedex-aws-spark", "client": "FedEx Digital International (formerly TNT Digital International)", "role": "Data Engineer — Freelance", "period": "2018 — 2020", "summary": "Worked in a 10–15 person data engineering and data science team inside the roughly 200-person former TNT Digital organisation, building and operating data workloads across AWS and GCP.", "stack": [ "AWS (EMR), GCP (GKE)", "BigQuery", "Terraform", "Kubernetes (kubectl + YAML)", "Spark, PySpark", "JupyterHub", "CI/CD" ], "outcomes": [ "Managed AWS and GCP platform infrastructure with Terraform and operated Spark/EMR ETL and JupyterHub workloads", "Developed, deployed and operated Spark workloads on an existing GKE cluster using Kubernetes YAML and kubectl", "Developed BigQuery workloads and productionised and performance-tuned Spark code written by data scientists", "Reworked one job covering 175 independent business cases: the old version OOMed after 3 hours without completing; the new version completed in under 5 minutes" ] }, { "slug": "abnamro-dial", "client": "ABN AMRO", "role": "Data Engineer — FTE", "period": "2018", "summary": "Contributed to the DIAL data platform build — consolidating fragmented departmental ETL onto a shared platform and migrating workloads from Hive to PySpark.", "stack": [ "Hive", "PySpark", "Spark", "Python" ], "outcomes": [ "Helped consolidate fragmented per-department ETL onto the shared DIAL platform", "Migrated workloads from Hive to PySpark", "Replaced a colleague's recurring 3-day-per-week manual Excel workflow with a 3-minute script" ] }, { "slug": "kpn-hortonworks", "client": "KPN", "role": "Big Data Consultant — FTE", "period": "2016 — 2018", "summary": "Hadoop cluster admin at KPN — installing and operating Hortonworks HDP clusters on bare-metal for external enterprise customers. Built the team's automated system-health framework.", "stack": [ "Hortonworks HDP", "Hadoop", "Bare-metal", "Robot Framework", "Python" ], "outcomes": [ "Operated Hortonworks HDP Hadoop clusters on bare-metal hardware for external enterprise customers", "Built an automated end-to-end system-health check suite in Robot Framework, replacing the manual install-verification routine", "Saved the team 8+ hours per week of repetitive manual verification" ] }, { "slug": "huawei-kpn-hlr-cutover", "client": "Huawei", "role": "Software Test Engineer — Telecom Core Database (HLR / HSS) — FTE", "period": "2009 — 2012", "summary": "Telecom core-network database work — HLR / HSS for mobile operators. Test & delivery lead on a major core-database cutover at KPN Netherlands (system capacity: 16M subscriber lines; 12M migrated; zero incidents).", "stack": [ "HLR", "HSS", "Telecom core network", "Test engineering", "Migration planning" ], "outcomes": [ "Test & delivery lead on the KPN NL core-network database cutover — system capacity of 16M subscriber lines, the most critical database in their core network", "~7 months of intensive pre-cutover testing; identified hundreds of bugs across all severity levels", "Led the design and execution of the migration plan, including staged rollouts and rollback procedures", "Zero incidents across months of cutover operations; 12M subscribers migrated successfully" ] } ] } }