Jaime Torres

Worked on Project Delorean, supporting the de-globalisation of legacy eBay/Shanghai analytical pipelines into Business Unit–owned data products, aligned with Adevinta’s decentralised data strategy.

Contributed across three Business Units, with a primary focus on Subito (Automobile.it) (~70%), and additional work for Benelux (BNL) and Kleinanzeigen (KA) (~30%).

Highlights:

- Led the migration of global ETLs into domain-owned, production-grade data products under a Data Mesh model.

- Rebuilt core Subito pipelines on Databricks, AWS S3/Athena, Airflow, Argo, Delta/Parquet, with strong governance, lineage, and SLA guarantees.

- Supported BNL with GCP → Databricks Unity Catalog ingestion patterns (BigQuery, GA4, PySpark, Airflow backfills).

- Contributed to privacy compliance initiatives in KA, building vendor-facing integrations for Data Deletion Requests (DDR) using Kotlin microservices and DynamoDB.

- Improved internal CDC/SCD engines, Airflow frameworks, and anonymisation tooling used across teams.

Contributed significantly to the GravityPlus Migration Project (Cloudera Exit), transitioning from Cloudera to Confluent to deliver real-time streaming services with sub-200ms latency for retail and corporate clients. The migration involved implementing a high-availability system across two data centers operating in parallel, with dual-region replication in the PaaS environment to ensure fault tolerance and reliability.

- Kafka Connect Jobs: Designed and implemented Kafka Connect jobs for data ingestion, preparation, and synchronization with Oracle, addressing challenges with complex data transformations.

- Enrichment Applications: Migrated 14 stateful streaming applications from Java 8 to Java 17 and Scala 2.13, ensuring seamless functionality and compliance with updated quality and security standards. Also migrated one legacy Apache Flume application to Kafka Streams streaming app.

- Historical Data Migration: Migrated 7 years of historical data from Cloudera Kafka topics and HBase, ensuring zero data loss and uninterrupted service during the transition.

- Documentation & Performance Analysis: Created comprehensive documentation, optimized system performance in production, and defined thresholds with alerting mechanisms for Kafka topic lag to enhance monitoring and operational efficiency.

🔸 Developed and launched of a comprehensive monitoring system for Europe DWH, enhancing real time ETL performance monitoring by 30% and reducing incident response times by 20%.

🔸 Collaborated in the design and implementation of a new global cloud-based data warehouse.

🔸 Established a simple incident management for Autos DWH and provided support for Autos data pipelines through global data services.