Confirmed on the employer's own hiring board on Nov 17, 2026. First seen by Alion on Oct 24, 2026.
Build AI agents that do the work, not just talk about it.
Neon Link places remote AI agent engineers in the Philippines with US companies. The newest AI tools don't stop at an answer: an agent reads a request, decides what to do, calls the right tools and finishes a multi-step job, with a person approving the steps that matter. You build those for US small and mid-size businesses: the agent itself, the tool connections that let it reach a client's systems safely, and the guardrails and approval points that keep it trustworthy. You also build the way the best engineers now build, with AI coding agents open every day, and you know where to trust them and where to check their work.
We care more about what you built than the courses you took: send us a link to an agent or tool server you built, tell us what it did and what kind of business ran it, and leave the client's name out. We are hiring for this role now. Depending on the engagement, the work is project based or full time, at a rate agreed up front. We never charge a fee to apply, test, train or start.
Pay: PHP 90,000 to 130,000 a month (about PHP 536 to 774 an hour). About USD 1,430 to 2,065 a month at today's exchange rate, for a full-time client engagement, depending on experience. Project work before that is paid per project, at a rate agreed up front.
Location: 100% remote, work from home. Open to applicants who live in the Philippines.
Setup: Remote, work from home, anywhere in the Philippines.
Engagement: Contract, project based or full time depending on the engagement, remote.
Schedule: Flexible, with at least 3 hours of overlap with US business hours (evening in the Philippines) for calls and handoffs.
Start: Hiring now.
What you'll do
- Find the multi-step job worth handing to an agent with the person who does it today, and map where a person must approve before the agent acts
- Build agents that take actions with tools, using the Claude Agent SDK, the OpenAI Agents SDK, LangGraph or plain API tool calling
- Build MCP servers and tool integrations that connect agents to the client's CRM, inbox, files, databases and internal APIs, with scoped permissions and an audit log of every call
- Design multi-step agent workflows with guardrails: input checks, step and spend limits, safe defaults, and human approval before anything reaches a customer, moves money or deletes data
- Use AI coding agents such as Claude Code and Cursor every day to build faster, and review and test what they write before it ships
- Evaluate agent reliability: task success on real cases, tool call errors, loops and stalls, with an eval suite you rerun on every change
- Write a runbook for each agent: what it can do, what it can't, who approves what, and how to switch it off
What success looks like
- Agents that finish the job on real cases, with the reliability bar and targets set with each client
- Nothing reaches a customer or a system of record without the approval the client asked for
- A client team that trusts the agent enough to use it every day, and knows how to pause it
What you bring
- 2 or more years shipping software in Python or TypeScript, including at least one agent that takes real actions with tools; a project of your own counts
- Hands-on with tool calling and at least one agent framework or SDK, such as the Claude Agent SDK, OpenAI Agents SDK or LangGraph
- Built or extended an MCP server or a comparable tool integration, and can explain how you handled auth and permissions
- Use AI coding agents such as Claude Code, Cursor or Codex daily, and can show how you review and test what they produce
- You measure agent reliability with evals and traces, not a few happy path demos
- Clear written English: you can explain what an agent will and won't do to a business owner who has never used one
- Careful with client data and access: you work only in the accounts a client sets up for you, keep their data, code and API keys out of personal accounts and public repos, and before your first client project you'll sign a confidentiality agreement and pass a background check, including a current NBI clearance
- Education: any degree or none; what you built matters more
Nice to have
- Browser or computer use agents, or workflow tools such as n8n running alongside agents
- A public repo, MCP server or agent project that other people use
Work setup
- Your own computer with 16 GB of RAM or more and an Intel Core i5, AMD Ryzen 5 or Apple M series processor or better, on Windows 10 or 11, macOS or Linux, with a webcam and a headset
- Fiber internet of at least 50 Mbps, with a mobile data or second provider backup
- A UPS or generator so a brownout doesn't drop a client call or a deploy
- A quiet workspace for client calls and screen shares
- Tell us your city and province when you apply; we hire across regions, so one storm or outage doesn't hit everyone at once
You'll do well here if
- Builds what's next
- Ships with guardrails
- Measures reliability honestly
- Explains it plainly
- Curious about the business
- Careful with other people's data
What happens next: A person looks at every link you send, and everyone who clears the first screen hears back, either way. If your work fits, you'll walk the founder through your agent on a call.
Your voice introduction
In your voice introduction (the form gives you up to three minutes), tell us who you are and explain the agent you linked the way you'd explain it to a US business owner: what it does on its own, where a person approves and how you'd switch it off. In the same recording, tell us how many hours a week you could give to paid projects now and when you could start full time.

