DevOps engineer turned AI agent engineer. At Tesote I own an internal agentic development platform: a fleet of 20 agent VMs that has opened 500+ PRs across 32 repositories, plus an hourly Linear sweep that has taken 400+ tickets end to end at about $2 of model spend per merged PR. Background in Linux, Terraform, Ansible, CI/CD and observability. C1+ English. Computer Science student at Universidad Nacional de Colombia.
Own the internal agentic development platform from architecture to production. Built the harness that turns VMs into AI workers: per-task templates (system prompt + model tier), pgvector retrieval over MCP, scoped credentials, and a full dev stack so agents can clone, migrate, boot the app and run tests unattended. Restructured repos for Claude Code with skills, scripts, hooks and MCP servers. Run Linux proxy servers on-prem and on DigitalOcean/Hetzner with Terraform and Coolify.
Logro principalAgents have opened 500+ PRs across 32 repos from a 20-VM fleet. An hourly Linear sweep has taken 400+ tickets end to end; 84% produced a PR, at about $2 of model spend per merged PR.
Administered the Azure tenant and deployed App Services and Logic Apps with Terraform and Bicep. Built CI/CD in Azure DevOps and GitHub Actions for containerized web apps, APIs and Function Apps (ACR, Container Apps). Automated cross-department workflows with Python, Logic Apps and n8n. Built data pipelines over SQL Server, Blob Storage, Salesforce and the Unicourt API with Data Factory, Python and stored procedures.
Logro principalSet up Terraform/Bicep deployments and CI/CD pipelines for the team's Azure web apps, APIs and Function Apps, and automated workflows across multiple departments with Python, Logic Apps and n8n.
Wrote Python and VBA scripts to automate manual, repetitive processes.
Logro principalSelected as 1 of 5 candidates from 60+ applicants for a Software Engineer apprenticeship based on technical aptitude.