Python Engineer · Data Systems & Automation · ETL · APIs · Reconciliation

I build systems for messy real-world data — where sources drift, inputs disagree, and a silent wrong answer is worse than a loud failure.

Open to: 100% remote Python / Data / Backend roles across LATAM and US-compatible teams · selected consulting work.

Portfolio · LinkedIn · Email

Reliable software is not just code that works on the happy path. I focus on systems that remain understandable and trustworthy when reality gets messy.

• Open-source adoption: chile-hub has grown into a practical public-data toolkit for Chile, with active community adoption and package distribution.
• Published Python packages: chile-hub, bankrecon and rutificador.
• Production ownership: I build and operate Tooltician, Monedario and Noticiencias.
• Engineering range: data pipelines, APIs, reconciliation, scraping, document processing, desktop tooling, CI/CD and TypeScript/Astro product work.

• chile-hub — turns fragmented Chilean public data into curated, normalized and validated datasets ready for Python, Polars, DuckDB, SQLite and Excel. _{Python · Data engineering · Open data · PyPI · CI/CD}

• conciliador_bancario — fail-closed bank reconciliation CLI built around deterministic matching, explicit ambiguity and audit-ready artifacts. Published as bankrecon. _{Python · Reconciliation · Financial workflows · Auditability · PyPI}

• rutificador — zero-dependency Python core plus CLI and integrations for validating, calculating and formatting Chilean RUT identifiers. _{Python · Validation · CLI · FastAPI/Pydantic integrations · PyPI}

• DNSpect — local-first DNS benchmarking across multiple protocols, designed to answer which resolver is actually better on your network rather than on somebody else's test server. _{Python · TypeScript · Networking · Benchmarking · Desktop packaging}

• PDF-Text-Analyzer — modular PDF ingestion, validation, multilingual extraction and search that keeps document failure modes visible instead of silently returning bad data. _{Python · PDF processing · NLP · Search · Batch pipelines}

I like owning the full feedback loop: build → ship → observe → correct → document.

Core: Python · SQL · FastAPI · Pydantic · Pandas · Polars · NumPy · Requests · Selenium · BeautifulSoup
Product: JavaScript/TypeScript · Astro · React · Node.js
Delivery: GitHub Actions · Docker · Linux · APIs · scheduled automation

• Make invalid states and uncertainty visible.
• Prefer deterministic behavior over clever but opaque heuristics.
• Treat observability, auditability and failure handling as product features.
• Keep contracts explicit between data, code and downstream consumers.
• Ship tests and documentation with the implementation.
• Optimize for software another engineer can understand, operate and extend.

If you are building data-heavy products, backend systems, financial workflows or automation where correctness under messy real-world conditions matters, I am interested in the problem.

Email me · Connect on LinkedIn · See the portfolio