I build fault-tolerant streaming pipelines, distributed compute workloads, and cloud warehouse infrastructure. My work centers on high-throughput event ingestion, in-memory stream processing, and automated, tested transformations.

Decoupled, high-throughput streaming architecture processing real-time telemetry events.

• Ingestion: Kafka broker with multi-partitioned topics for parallel event consumption.
• Stream Processing: PySpark Structured Streaming for in-memory anomaly detection and noise reduction.
• Storage & Orchestration: PostgreSQL intermediate buffer with idempotent loads to Snowflake via Mage AI.
• Modeling & Testing: Staging and dimensional star schema layers managed and validated with dbt Core.
• Environment: Fully containerized via Docker Compose with volume-backed persistence.

• Decoupled Architecture: Separating ingestion, compute, and storage layers to prevent cascading pipeline failures.
• Idempotency by Design: Ensuring every DAG and micro-batch run can be re-executed without duplicating or corrupting target warehouse state.
• Schema Integrity: Enforcing strict type checks, null constraints, and automated dbt tests upstream before data hits downstream consumers.

• Upwork: Available for Data Engineering, dbt modeling, and streaming infrastructure contracts.
• Code: Check my pinned repositories below for architecture diagrams and production codebases.