I build automation systems for marketing agencies and e-commerce brands.

Problem: Agüita, the stainless-steel bottle brand I acquired, ran, and later sold in Tenerife (physical stock, B2B wholesale, in-house laser engraving), had me as its reporting department. Every week I pulled Shopify and GA4 numbers by hand to see what was selling and where the store stood: two to three hours of it, every week. First job I fired myself from.

System: Make scenario pulls Shopify + GA4 data nightly → writes to a Google Sheets master → digest lands in Slack every morning, summary every Monday at 7am. Error alerts on failure.

Result: Reporting ran itself from launch until the day I sold the brand. The same brand also ran an automated revenue-split report for a collab: tagged-SKU sales, extracted and delivered to the partner, another three to four hours a month I stopped spending. I was the test bench.

Problem: While I was a contracted engineer inside TIM ($16B revenue), my team was burning 15 hours a week on the same data-crunch: pulling, cleaning, reconciling, exporting. Manual every Monday.

System: Scheduled Bash cron → Python scripts pull from source systems → clean, reconcile, validate → write outputs in the same downstream format. Logs to disk. Failure alerts before anyone notices.

Result: 15 hours/week removed. ~60 hours/month of senior analyst time recovered. Monday bottleneck eliminated.

Problem: My own drone-video side business (real-estate videos, commercials, documentaries) was landing 2 client leads a month, all from word of mouth. I was flying shoots all day, zero time left for prospecting. So I became my own first lead-gen client.

System: Apollo scrapes and filters real-estate agencies → OpenAI writes a personalized opening line per prospect → Mailgun sends the sequence with reply detection → replies land in Pipedrive, tagged and ready. Built in 30 days.

Result: Monthly client leads tripled (2 → 6) in the first 30 days. My only job left in the loop: answering warm replies.