Abhishek singh Baghel

I build AI automation that does real work without an operator, and I measure it. Since March 2025 I have worked as a self-employed AI Automation Engineer: 18 AI content pipelines I designed run unattended every day, each taking a video from research to published in about an hour of compute. I design the process, the rules and the prompts, direct Claude Code to implement them, and read the running system to catch faults. Two results: - One channel went from 4 to 5 hours of manual editing a day to publishing on 6 to 7 channels a day. The reliability layer behind it has 151 rules, 82 automatic gates and 146 tests, built from 67 real production failures. - My open source customer-message triage system routed 25 of 25 test messages correctly in four consecutive live runs, passed a 59-attack prompt-injection suite with no harmful reply in the last four runs, and costs $0.35 per 1,000 messages. My public n8n portfolio (a WhatsApp lead qualification agent, and error handling with a failure ledger) ships every workflow with a labelled evaluation and its limits written down. Before this I ran a fully remote content agency for SaaS and app clients from June 2022 to December 2025, led a team of four editors, and built and sold four content properties. Next I want a remote AI automation role where I own workflows end to end: n8n, LLM agents, CRM and messaging integrations, with results I can measure.