I design cloud data platforms and ML pipelines that transform raw events into reliable, observable and decision-ready products. My work combines data engineering foundations with MLOps and AI capabilities: ingestion, streaming, lakehouse storage, orchestration, quality, lineage, model lifecycle and monitoring.
| ๐ Location Italy |
โฑ๏ธ Experience 3+ years |
๐ Education Artificial Intelligence Engineering |
| โ๏ธ Cloud AWS ยท Databricks ยท S3 |
โก Data Spark ยท Kafka ยท Airflow ยท SQL |
๐ค ML/AI MLOps ยท MLflow ยท Computer Vision |
| Domain | Focus |
|---|---|
| โ๏ธ Cloud Data Platforms | AWS, Databricks, S3, Snowflake, lakehouse and scalable storage |
| ๐ Streaming & CDC | Kafka, Debezium, Flink, Spark Structured Streaming, replay and idempotence |
| ๐งฑ ETL / ELT | Python, SQL, Airflow, dbt, Bronze/Silver/Gold and data marts |
| ๐ค MLOps & AI | MLflow, Feast, Evidently, feature engineering, anomaly detection and model serving |
| โ Data Quality & Governance | Great Expectations, DataHub, Avro contracts, schema evolution and lineage |
| ๐ Reliability & Observability | Prometheus, Grafana, OpenLineage, alerting, checkpoints and recovery strategies |
01 ยท CDC Lakehouse NYC
CDC and analytics architecture for NYC taxi data:
NYC TLC โ PostgreSQL โ Debezium โ Kafka โ Spark Structured Streaming
โ Apache Hudi on S3 โ Snowflake โ dbt marts โ BI & analytics
Key themes: replayability ยท idempotent upserts ยท schema contracts ยท data quality ยท lineage ยท observability
02 ยท E-Commerce Data Platform
Modern data platform combining real-time and batch processing with Kafka, Flink, Spark, Delta Lake, Snowflake, dbt, Airflow, MLflow, DataHub and Great Expectations.
Industrial IoT pipeline connecting PLC/Modbus telemetry to MQTT, Kafka, InfluxDB, PostgreSQL and Grafana, with edge buffering, monitoring and anomaly detection.
Digital-twin architecture built around Unified Namespace, edge intelligence, AI-assisted operations, energy/COโ monitoring, FastAPI, InfluxDB and Grafana.
Anomaly-detection system using truncated/randomized SVD, residual scoring, EWMA alerting, Redis state, MLflow tracking, Kafka and Prometheus/Grafana.
AI interaction prototype combining gesture recognition, speech commands, facial-expression analysis and multimodal decision rules with Python and OpenCV.
| Institution | Program | Period |
|---|---|---|
| Universitร degli Studi di Urbino Carlo Bo | Engineering Degree โ Artificial Intelligence | September 2025 โ Present |
| Universidad Complutense de Madrid | Engineering Degree โ Artificial Intelligence | September 2026 โ June 2027 |
- Architecture with explicit reliability, quality and recovery guarantees
- Data products designed for replay, scale and operational visibility
- MLOps thinking: experiment tracking, feature management, drift and monitoring
- Documentation that explains decisions, trade-offs and maturity honestly
- A pragmatic delivery style: build incrementally, validate assumptions and measure results
๐ Portfolio ยท ๐ผ LinkedIn ยท ๐ง a.briki@campus.uniurb.it
Building reliable data systems, one pipeline at a time.