# Data Engineering Projects

> Explore Mario Caesar's data engineering projects across data quality, SQL tooling, orchestration, warehouses, and production-minded pipeline design.

Canonical page: https://caesarmar.io/projects

## SQLTerrain - Local-First SQL Reasoning Workbench

Personal Project | Jun 2026 - Present

SQLTerrain is a local-first SQL workbench for semantic query analysis, cross-dialect conversion, local data profiling, and guided query building with Next.js, SQLGlot/Pyodide, and DuckDB-WASM. I tested its behavior against 10,000 Atlas benchmark cases across 12 semantic dimensions and 1,224 SQLGlot conversion tests, with explicit fallback and provenance handling.

Technologies: SQL, SQLGlot, DuckDB-WASM, Data Profiling

[Project source](https://github.com/caesarmario/sqlterrain-showcase)

## Agentic Data Quality Triage

Personal Project | May 2026 - Present

This local data quality platform runs against synthetic order data. It checks the warehouse, collects pipeline and lineage evidence after a failure, and prepares a report for an operator to review before approving a backfill.

Technologies: Data Quality, Airflow, dbt, ClickHouse, LangGraph

[Project source](https://github.com/caesarmario/agentic-data-quality-triage)

## Automating Kaggle Stock Dataset Updates

Personal Project | Dec 2025 - Jan 2026

This stock data pipeline refreshes Yahoo Finance data and publishes new dataset versions through the Kaggle API. Full-refresh loads are idempotent, overwrites are atomic, and each run records data-quality checks in a JSON receipt.

Technologies: Google Colab, Pandas, Kaggle API, Requests

[Project source](https://www.kaggle.com/code/caesarmario/auto-update-existing-kaggle-datasets-via-api/)

## Config-Driven Yelp and Weather Data Quality Pipeline

Personal Project | Jun 2025 - Jul 2025

This configurable pipeline handles Yelp and weather data, including large source files and quality checks across staging, L1, and warehouse layers.

Technologies: Airflow 3, Pandas, Config-Driven ETL, Data Quality

[Project source](https://github.com/caesarmario/config-driven-dq-yelp-weather)

## Loan Applicant Data Warehouse (Pentaho)

Academic Project | Jan 2021 - Jan 2022

Modeled raw loan-applicant records into clean warehouse tables for reporting, analysis, and dashboards.

Technologies: Pentaho ETL, Dimensional Modeling, BI Reporting

[Project source](https://github.com/caesarmario/data-warehouse-credit-card-applicant-using-pentaho)