Senior Engineer for AI Agent Automation (Claude Code / Codex, CI, TypeScript, Python or Go)

Please don't send a generic proposal. I read every proposal myself, and copy-paste or AI-written templates are declined right away. If you don't use coding agents hands-on in your daily work, please save your Connects. I'm a senior engineer building AI agent automation for an engineering team, and I need a second senior developer to build alongside me. The work is practical. Coding agents run on a schedule or on events. They pick up tickets, reproduce bugs, open pull requests and attach proof that the change works. Everything around them is ordinary code: integrations with an issue tracker and GitHub, CI checks, test runs that record evidence, and logs of what each run did. Some of it is new territory, so you'll often be learning a tool in the morning and shipping with it by the end of the week. WHAT YOU'LL DO - Build agent workflows that run without a person driving them (headless coding agents, hooks, MCP servers, custom skills or commands) - Write the plumbing in TypeScript, Python or Go: API integrations, sync jobs, CLI tools - Set up GitHub Actions jobs, required checks and test runs that leave clear evidence - Find a tool's limits quickly, test them, and write down what you found - Keep everything in the repo with short runbooks so others can run it YOU SHOULD HAVE - Senior-level, hands-on engineering experience: you've built and shipped real systems yourself - Daily use of coding agents (Claude Code, Codex, Cursor or similar) in your own work - Solid skills in at least one of TypeScript, Python or Go, plus Git and GitHub - The ability to pick up an unfamiliar tool fast and get something working with it - Clear written and spoken English and the ability to work independently from a short brief NICE TO HAVE - Running coding agents outside of chat, in scripts, CI or on a schedule - Building MCP servers, agent hooks or custom skills - GitHub Actions and PostgreSQL - Playwright or other browser automation - Workflow orchestrators (Kestra, Airflow, Temporal, Prefect) - Jira or other issue tracker APIs - OpenTelemetry ENGAGEMENT - Ongoing, hourly, 25-30 hrs/week to start - Remote. Your working day must overlap US Eastern business hours through 5 PM ET - Private repos and NDA. No production data or customer personal data is ever shared with AI tools. TO APPLY (REQUIRED) - Start your proposal with the word "HANDS-ON". - Include a Loom link showing a project you built running. Use a Loom you already have; please don't record a new one for this post. - The project can be anything real you built: an agent, an automation, a backend service, a data pipeline, a dev tool, a web app. Agent or automation work is a plus but not required. Client work is fine if you blur anything sensitive. - Proposals without a Loom link won't be considered. - Please keep screening answers short (1-2 sentences each). HIRING PROCESS 1. Review: I read the proposals and watch the Looms. 2. Video call: shortlisted candidates join a call with the camera on. You'll share your screen and walk me through a past project: open the code, run it, and explain how it's built. Pick something you know well, because I'll ask questions as we go. 3. Paid trial: a small, self-contained task (about 3-4 hours, paid at your hourly rate) using public code, then a short call where you show and explain what you built. 4. Hire.