OMG!! NEEDED NOW!! Agentic Operations Engineer – Paperclip Agent Fleet, n8n & Python Infra – Fast Growing ETL Agency

Hey there!

I'm looking for ONE HIGHLY TEACHABLE (Willing to Learn) Agentic Operations Engineer with excellent written English to run and scale our Paperclip agent fleet.

We run a workforce of AI agents in Paperclip. They have roles, reporting lines, goals, budgets, and assigned tasks, exactly like the human side of the company. They read our systems, classify what comes in, route work to the right person, and make low-risk decisions on their own inside hard guardrails. You will hire agents into that fleet, give them goals, set and defend their budgets, watch what they cost per task, pause the ones that drift, and close out budget i ---------- when a scope hard-stops.

This is agentic operations, not prompt engineering. The prompts are the easy part. The job is owning a fleet that acts without supervision and making sure it never quietly does damage.

One fixed part-time shift. It can grow into a full-time role where you own the entire agent fleet if you prove to be a great fit.

Watch this video to learn more about how our agency works and to also see my personality and style:
---------- /share/b301319feb524d55acb8eacb631cb7c1?sid=b843a7e2-95d ---------- a0-51b42598f74f

ABOUT YOU:
Great communicator in writing
Won't be easily offended
Extremely reliable
Positive, no-excuse attitude!
Will stop and ask when something is unclear instead of guessing
Deeply suspicious of confident output that nobody checked

KNOWLEDGE REQUIRED:
Real production debugging, meaning you have diagnosed and fixed things that were broken in live systems using logs, traces, and configuration checks, not only local development issues
Hands-on with agentic systems, meaning you have run something that took actions on its own rather than only answering questions, and you have dealt with what happens when it takes the wrong one
Working directly with LLM APIs and MCP tool surfaces in production, including tool call failures, retries, and bad output handling
Designing guardrails: allow lists, hard stops, spend caps, kill switches, and audit trails, so an autonomous system cannot act outside its scope
Cost control over agent runs, tracking spend per task against a budget and knowing why a number moved
Python, ideally FastAPI or a similar API framework
n8n, building, maintaining, and troubleshooting workflows, schedules, and integrations
Proficient in Linux Bash commands
Docker, building and troubleshooting container images, tags, runtime configuration, and application startup failures
Database troubleshooting and writing SQL against a warehouse to verify what a system actually did, rather than trusting what it reported
Kubernetes basics, diagnosing pod failures, deployments, and configuration problems
Comfortable with Codex, Claude Code, or similar AI-assisted tools, while independently reviewing, validating, and correcting what they produce
GoHighLevel APIs, OAuth credentials, and integration troubleshooting is a strong plus
Familiarity with Ubuntu, ProxMox, and Git-based deployment workflows

ABOUT ME (Alex):
I'm a CEO with over 20 years of experience in digital marketing
I move very fast & expect the people I work with to be the same way
I'm very loyal to my employees & treat them with respect
I ALWAYS give bonuses for doing a good job (if I make money, you should be rewarded for that)
I run the company remotely and am located in Indonesia

ABOUT THE POSITION:
You own the agent fleet end to end. Day to day that means hiring agents into Paperclip and defining their roles and reporting lines, writing the goals they work toward, assigning and tracking their tasks, setting budgets and watching observed spend against them, resolving budget i ---------- when a scope hard-stops, and pausing an agent or a whole company when something goes wrong. Around that you build the plumbing: scheduled agents in n8n that read our internal systems and act on what they find, wired to model APIs, with the audit trail that proves what each one did and why.

The part of this job I care about most is judgment, and in agentic work it matters more than anywhere else in the company. An agent that is confidently wrong is worse than an agent that does nothing, because it moves fast, it looks correct, and it keeps going. I need someone whose instinct is to go and check what the system actually did rather than trust what the output says it did. When a task is unclear, when there is a risk nobody has named, or when you are missing context you need, I want you to stop, lay out the options and their tradeoffs, and ask for direction. I do not want a confident guess that quietly breaks something a week later. Engineers who ask the right question at the right ---------- nt are worth far more to me than engineers who never ask anything.

Schedule, one seat, four hours per shift, five days a week:
Morning seat, 8:00 AM to 12:00 PM Manila time
Evening seat, 8:00 PM to 12:00 AM Manila time
Tell me which one you want.

ABOUT THE COMPENSATION:
We offer DAILY PAYMENTS on a task-by-task basis through WISE, ensuring timely and efficient compensation for your work.
We pay bonuses for great work. We reward results. This job is about providing value.
13th-month bonus for candidates who qualify for a full-time position, which includes ongoing monitoring and improvement of our systems.

HOW TO APPLY:
Send an email with the subject line: Awesome DevOps Engineer
Include your resume
Tell me which shift you want, morning or evening
Tell me about an agentic or automated system you built that ran without you watching it, and what it did wrong before you caught it. I want the specific way you found out it was wrong, not the fix.
Then tell me: if you were running a fleet of agents with real budgets and real access, what is the first thing you would refuse to let them do on their own, and why?

Congratulations! You made it to the end, so what are you waiting for?
Please email me at: ----------