I architected and deployed a logistics intelligence platform as an independent AI Systems Engineer, using Python, n8n, OpenAI APIs, and multimodal ingestion. It processed 200,000+ operational emails across 700+ projects and automated ~70% of operational tracking.
I built structured records from email threads, PDFs, and OCR-derived content, then added traceability for shipment routes, artwork status, and artist histories. I also implemented checkpointing, recovery, and backup logic to preserve operational knowledge.
For my voice activity detection thesis at UNAM, Faculty of Sciences, I designed and trained five neural architectures and ran 90+ controlled experiments. The work showed how model performance differed across noise conditions and operating regimes, even when validation accuracy was identical.
At Smart Solutions, I delivered data analysis and validation for clients including HSBC and Qualitas Insurance, and contributed to workflow changes that accelerated project delivery by ~50%. At Atelier d'Art, I designed communication and lead-conversion systems that helped grow enrollment from 20 to 110.

