Automation
CLI-first tools, internal workflows, structured execution, state, and validation.
Oscar Gracia
I build practical Python systems for workflow automation, document retrieval, and reviewable AI-assisted work. My strongest proof is project-backed: orchestration tooling, CLI-first AI utilities, retrieval pipelines, and monitoring-oriented analytical systems.
Focus
My current positioning is centered on Python automation, applied AI, document workflows, and analytical infrastructure. I work best on systems that need inspectability, repeatable execution, and clear operational value.
CLI-first tools, internal workflows, structured execution, state, and validation.
Responses API integrations, structured outputs, retrieval flows, and grounded review.
Market-data processing, alerting, monitoring, and operator-facing reporting.
Selected Work
Proposed multi-agent coding system that routes routine implementation to cheaper models while enforcing budgets, retries, validation, Git checkpoints, and escalation.
Technical note on constrained agent design: deterministic software for rules and validation, focused LLM use for ambiguous language and reasoning tasks.
Python CLI workflow for finding, enriching, scoring, and reviewing business leads before manual outreach, with structured OpenAI analysis only on qualified rows.
Frontend-only demo of a custom reporting workflow that cleans exported marketing CSVs, flags anomalies, and drafts client-ready KPI summaries for human review.
Terminal-first Python orchestration for bounded coding-task workflows with planning, execution, review, and resumable state.
Document-ingestion and retrieval workflow with metadata persistence, validation, and reviewable question-answering over indexed legal materials.
Lightweight CLI utility around the OpenAI Responses API with local-file input, attachments, and retrieval-oriented document workflows.
Python-based analytical infrastructure for market-data ingestion, time-series processing, monitoring, charting, and operator-facing alerts.
Best Fit
The strongest fit is technical work where implementation quality matters more than hype: workflow automation, retrieval systems, AI evaluation, technical operations, and documentation-heavy problem spaces.
Contact
If you need Python automation, retrieval workflows, or structured AI review support, email is the fastest way to reach me.